Add Qwen3-Omni held-out error analysis

ARTIFACT_GUIDE.md

@@ -110,12 +110,14 @@ research project.
 | [`results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md`](results/omni_finetune/MULTI_EPISODE_ACCESS_STATUS.md) | Documents the public multi-episode access path, selected 128-episode pilot plan, and data requirements. |
 | [`docs/data/omni_finetune_verified_result.json`](docs/data/omni_finetune_verified_result.json) | Compact verified summary for the first selected-episode Qwen3-Omni diagnostic pilot, including split counts, held-out metrics, and the quality-target caveat. |
 | [`results/omni_finetune/verified_public/`](results/omni_finetune/verified_public/) | Public-safe verified held-out result packages. These include metrics, predictions, reports, manifests, training metadata, validation summaries, and audit files, but not raw data or weights. |
+| [`results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md`](results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md) | Derived held-out error analysis by episode, action family, train-seen status, required-modality state, and object category for the validation-aware Qwen3-Omni diagnostic pilot. |
 | [`scripts/omni/discover_xperience10m_sources.py`](scripts/omni/discover_xperience10m_sources.py) | Discovery gate for valid multi-episode Xperience-10M sources. |
 | [`scripts/omni/train_qwen3_omni_lora.py`](scripts/omni/train_qwen3_omni_lora.py) | Training entrypoint for the Qwen3-Omni LoRA pilot after the data gate passes. |
 | [`scripts/omni/run_128_fullsplit_parallel_export_8gpu.sh`](scripts/omni/run_128_fullsplit_parallel_export_8gpu.sh) | Full 96/16/16 launcher with parallel export, 8-process LoRA training, validation-sample monitoring, held-out test evaluation, and quality-target reporting. |
 | [`scripts/omni/merge_qwen3_omni_eval_shards.py`](scripts/omni/merge_qwen3_omni_eval_shards.py) | Recomputes held-out metrics from deterministic Qwen eval shards and checks missing or duplicate prediction ids. |
 | [`scripts/omni/package_verified_omni_result.py`](scripts/omni/package_verified_omni_result.py) | Creates a contract-driven public-safe package from validated held-out fine-tuning outputs without raw data, base weights, adapter/checkpoint weights, full checkpoints, or large archives. |
 | [`scripts/omni/audit_verified_omni_package.py`](scripts/omni/audit_verified_omni_package.py) | Audits a verified package before README, website, or Hugging Face updates by checking validation status, required files, primary metrics, held-out evidence, and forbidden file types. |
+| [`scripts/omni/analyze_qwen3_omni_errors.py`](scripts/omni/analyze_qwen3_omni_errors.py) | Computes public-safe held-out error-analysis tables from the verified Qwen3-Omni prediction package. |
 | [`scripts/omni/watch_verified_omni_package.py`](scripts/omni/watch_verified_omni_package.py) | Waits for a passing held-out eval validation and then runs the verified public-safe packager automatically. |
 | [`OMNI_MODEL_EXTENSION_CONTRACT.md`](OMNI_MODEL_EXTENSION_CONTRACT.md) | Human-readable contract for adding new model families while preserving the same episode split, held-out evaluation, packaging gate, and public-safety boundary. |
 | [`configs/omni_backbones/`](configs/omni_backbones/) | Backbone registry for implemented Qwen3-Omni LoRA plus planned Cosmos-style world-model and VLA/policy branches. |

PROJECT_STATUS.md

@@ -30,7 +30,7 @@ scale-up readiness; it is not presented as final full-dataset model quality.
 | Public dashboard and Hub pages | Verified | GitHub Pages, HF Space, artifact dataset, baseline model repo, Qwen3-Omni LoRA repo | Readers can move between the website, code, derived artifacts, baseline weights, and Qwen3-Omni pilot status without needing local infrastructure details. |
 | Public package policy | Verified | `DATA_NOTICE.md`, `REPRODUCIBILITY.md` | Raw Xperience-10M data, private gated files, large archives, credentials, and full Qwen weights are not redistributed. |
 | Reproducibility | Verified for the public sample | `REPRODUCIBILITY.md`, `docs/data/reproducibility_matrix.json`, `notes/reproducibility_audit.md` | The public sample workflow has explicit commands, expected outputs, and exact-match reproduction evidence. |
-| Qwen3-Omni fine-tuning | Verified validation-aware diagnostic held-out pilot; quality target not met | `docs/data/omni_finetune_verified_result.json`, `results/omni_finetune/verified_public/`, `scripts/omni/package_verified_omni_result.py`, `scripts/omni/audit_verified_omni_package.py` | The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, and 448 test predictions. JSON validity is 87.50%, below the 98% target, so the result is a diagnostic baseline and the next pass should focus on structured-output improvements and error analysis. |
+| Qwen3-Omni fine-tuning | Verified validation-aware diagnostic held-out pilot; quality target not met | `docs/data/omni_finetune_verified_result.json`, `results/omni_finetune/verified_public/`, `results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/`, `scripts/omni/package_verified_omni_result.py`, `scripts/omni/audit_verified_omni_package.py`, `scripts/omni/analyze_qwen3_omni_errors.py` | The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, 448 test predictions, and derived error-analysis tables by episode, action family, train-seen status, required-modality state, and object category. JSON validity is 87.50%, below the 98% target, so the result is a diagnostic baseline and the next pass should focus on structured-output improvements. |
 | Raw Xperience-10M redistribution | Not included | `DATA_NOTICE.md`, `docs/data/publication_audit.json` | Raw MP4, HDF5, RRD files, private gated data, and full Qwen weights are intentionally excluded. |
 
 ## Fast Research Route

data/artifact_index.json

@@ -1,12 +1,12 @@
 {
   "title": "Ropedia Xperience-10M Task Suite Artifact Index",
-  "generated_at_utc": "2026-06-06T14:35:42+00:00",
+  "generated_at_utc": "2026-06-06T14:53:45+00:00",
   "status": "pass",
-  "artifact_count": 83,
+  "artifact_count": 86,
   "missing": [],
   "by_kind": {
     "project_path": 14,
-    "scaleup_contract": 6,
+    "scaleup_contract": 7,
     "project_scope": 1,
     "source_alignment": 5,
     "publication_workflow": 3,
@@ -28,7 +28,7 @@
     "onboarding_doc": 1,
     "generated_figure": 3,
     "generated_figure_assets": 1,
-    "scaleup_status": 2,
+    "scaleup_status": 4,
     "citation": 1,
     "license": 1
   },
@@ -63,8 +63,8 @@
       "surface": "repo_hf",
       "shows": "Gives a compact current-state table for first-pass readers.",
       "exists": true,
-      "bytes": 8534,
-      "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+      "bytes": 8805,
+      "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
     },
     {
       "id": "project_status_json",
@@ -74,8 +74,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable copy of the current project status for website and HF mirrors.",
       "exists": true,
-      "bytes": 10977,
-      "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+      "bytes": 11274,
+      "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
     },
     {
       "id": "research_roadmap",
@@ -187,6 +187,17 @@
       "bytes": 6519,
       "sha256": "a3773fc681e298325e2be80556d6be6e7e30b90ba22ee24b66633f07ff9c4ea4"
     },
+    {
+      "id": "qwen3_omni_error_analysis_script",
+      "title": "Qwen3-Omni held-out error-analysis script",
+      "path": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "kind": "scaleup_contract",
+      "surface": "repo_hf",
+      "shows": "Computes public-safe held-out error-analysis tables by episode, action family, train-seen status, required-modality state, and object category.",
+      "exists": true,
+      "bytes": 15676,
+      "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+    },
     {
       "id": "additional_development_directions",
       "title": "Additional development directions",
@@ -250,8 +261,8 @@
       "surface": "repo_hf",
       "shows": "Gives the human-readable map from project scope to data, tasks, platform mirrors, and scale-up status.",
       "exists": true,
-      "bytes": 15660,
-      "sha256": "a9ad335b82c35a5ac102428663ffae1c8798e90e45cc5e795c3a499b4563b417"
+      "bytes": 16318,
+      "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
     },
     {
       "id": "official_dataset_card_alignment",
@@ -695,8 +706,8 @@
       "surface": "repo_hf",
       "shows": "Generates the selective artifact catalog from local files.",
       "exists": true,
-      "bytes": 30785,
-      "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+      "bytes": 32191,
+      "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
     },
     {
       "id": "publication_audit",
@@ -731,7 +742,7 @@
       "volatile": true,
       "shows": "Confirms prepared GitHub/HF Space/artifact/model mirrors share the same critical data, figure, website HTML, and validator files.",
       "exists": true,
-      "bytes": 111950,
+      "bytes": 126335,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -933,6 +944,28 @@
       "bytes": 3076,
       "sha256": "23b87581cfc1d95b0af118a0dbb4e601f42fc6bad608759490e13a9a1ef73205"
     },
+    {
+      "id": "qwen3_omni_error_analysis_report",
+      "title": "Qwen3-Omni held-out error-analysis report",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "kind": "scaleup_status",
+      "surface": "repo_hf",
+      "shows": "Summarizes validation-aware Qwen3-Omni held-out failures by episode, action family, train-seen status, required-modality state, and object category.",
+      "exists": true,
+      "bytes": 3331,
+      "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+    },
+    {
+      "id": "qwen3_omni_error_analysis_json",
+      "title": "Qwen3-Omni held-out error-analysis JSON",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "kind": "scaleup_status",
+      "surface": "repo_hf",
+      "shows": "Machine-readable Qwen3-Omni held-out error analysis with grouped metrics and sanitized failure examples.",
+      "exists": true,
+      "bytes": 25202,
+      "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+    },
     {
       "id": "citation",
       "title": "Citation metadata",

data/mirror_parity.json

@@ -1,16 +1,20 @@
 {
-  "status": "pass",
-  "generated_at_utc": "2026-06-06T14:37:36+00:00",
+  "status": "fail",
+  "generated_at_utc": "2026-06-06T14:55:21+00:00",
   "hf_root": "hf_publish",
   "summary": {
-    "group_count": 104,
-    "failure_count": 0,
-    "failures_by_surface": {}
+    "group_count": 114,
+    "failure_count": 32,
+    "failures_by_surface": {
+      "hf_space": 10,
+      "hf_artifacts": 11,
+      "hf_model": 11
+    }
   },
   "checks": [
     {
       "name": "repo_hf_space_artifact_model_data_parity",
-      "status": "pass"
+      "status": "fail"
     },
     {
       "name": "repo_hf_visual_asset_parity",
@@ -18,7 +22,7 @@
     },
     {
       "name": "repo_hf_validator_script_parity",
-      "status": "pass"
+      "status": "fail"
     },
     {
       "name": "repo_hf_website_html_parity",
@@ -26,7 +30,7 @@
     },
     {
       "name": "repo_hf_diagnostic_result_parity",
-      "status": "pass"
+      "status": "fail"
     },
     {
       "name": "repo_hf_quality_doc_parity",
@@ -98,34 +102,56 @@
     },
     {
       "name": "data/artifact_index.json",
-      "status": "pass",
+      "status": "fail",
       "local": {
         "path": "repo:docs/data/artifact_index.json",
         "exists": true,
-        "bytes": 37736,
-        "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+        "bytes": 39486,
+        "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
         },
         "hf_model": {
           "path": "hf_model:metrics/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
         }
       },
-      "failures": []
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:data/artifact_index.json",
+          "expected_sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0",
+          "actual_sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:docs/data/artifact_index.json",
+          "expected_sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0",
+          "actual_sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:metrics/artifact_index.json",
+          "expected_sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0",
+          "actual_sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
+        }
+      ]
     },
     {
       "name": "data/brand_assets.json",
@@ -350,27 +376,27 @@
       "local": {
         "path": "repo:docs/data/omni_finetune_verified_result.json",
         "exists": true,
-        "bytes": 3145,
-        "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+        "bytes": 4142,
+        "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         },
         "hf_model": {
           "path": "hf_model:metrics/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         }
       },
       "failures": []
@@ -474,61 +500,83 @@
       "local": {
         "path": "repo:docs/data/project_status.json",
         "exists": true,
-        "bytes": 10977,
-        "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+        "bytes": 11274,
+        "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         },
         "hf_model": {
           "path": "hf_model:metrics/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         }
       },
       "failures": []
     },
     {
       "name": "data/publication_audit.json",
-      "status": "pass",
+      "status": "fail",
       "local": {
         "path": "repo:docs/data/publication_audit.json",
         "exists": true,
         "bytes": 7237,
-        "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+        "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
         },
         "hf_model": {
           "path": "hf_model:metrics/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
         }
       },
-      "failures": []
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:data/publication_audit.json",
+          "expected_sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d",
+          "actual_sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:docs/data/publication_audit.json",
+          "expected_sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d",
+          "actual_sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:metrics/publication_audit.json",
+          "expected_sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d",
+          "actual_sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
+        }
+      ]
     },
     {
       "name": "data/public_surface_qa.json",
@@ -811,34 +859,56 @@
     },
     {
       "name": "data/scope_claims_audit.json",
-      "status": "pass",
+      "status": "fail",
       "local": {
         "path": "repo:docs/data/scope_claims_audit.json",
         "exists": true,
         "bytes": 20823,
-        "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+        "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
         },
         "hf_model": {
           "path": "hf_model:metrics/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
         }
       },
-      "failures": []
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:data/scope_claims_audit.json",
+          "expected_sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3",
+          "actual_sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:docs/data/scope_claims_audit.json",
+          "expected_sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3",
+          "actual_sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:metrics/scope_claims_audit.json",
+          "expected_sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3",
+          "actual_sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
+        }
+      ]
     },
     {
       "name": "data/single_episode_explorer.json",
@@ -935,34 +1005,56 @@
     },
     {
       "name": "data/task_surface_integrity.json",
-      "status": "pass",
+      "status": "fail",
       "local": {
         "path": "repo:docs/data/task_surface_integrity.json",
         "exists": true,
         "bytes": 45779,
-        "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+        "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
         },
         "hf_model": {
           "path": "hf_model:metrics/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
         }
       },
-      "failures": []
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:data/task_surface_integrity.json",
+          "expected_sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6",
+          "actual_sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:docs/data/task_surface_integrity.json",
+          "expected_sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6",
+          "actual_sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:metrics/task_surface_integrity.json",
+          "expected_sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6",
+          "actual_sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
+        }
+      ]
     },
     {
       "name": "data/task_walkthroughs.json",
@@ -997,34 +1089,56 @@
     },
     {
       "name": "data/website_integrity.json",
-      "status": "pass",
+      "status": "fail",
       "local": {
         "path": "repo:docs/data/website_integrity.json",
         "exists": true,
         "bytes": 15221,
-        "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+        "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
         },
         "hf_model": {
           "path": "hf_model:metrics/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
         }
       },
-      "failures": []
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:data/website_integrity.json",
+          "expected_sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2",
+          "actual_sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:docs/data/website_integrity.json",
+          "expected_sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2",
+          "actual_sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:metrics/website_integrity.json",
+          "expected_sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2",
+          "actual_sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
+        }
+      ]
     },
     {
       "name": "data/xperience10m_dataset_card_alignment.json",
@@ -1723,6 +1837,46 @@
       },
       "failures": []
     },
+    {
+      "name": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "status": "fail",
+      "local": {
+        "path": "repo:scripts/omni/analyze_qwen3_omni_errors.py",
+        "exists": true,
+        "bytes": 15676,
+        "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+      },
+      "mirrors": {
+        "hf_artifacts": {
+          "path": "hf_artifacts:scripts/omni/analyze_qwen3_omni_errors.py",
+          "exists": true,
+          "bytes": 15655,
+          "sha256": "e90ffd4bb75b001ab41cd956dfbb0a99b574d0b5e8ffc1a64e2887490d658daa"
+        },
+        "hf_model": {
+          "path": "hf_model:scripts/omni/analyze_qwen3_omni_errors.py",
+          "exists": true,
+          "bytes": 15655,
+          "sha256": "e90ffd4bb75b001ab41cd956dfbb0a99b574d0b5e8ffc1a64e2887490d658daa"
+        }
+      },
+      "failures": [
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:scripts/omni/analyze_qwen3_omni_errors.py",
+          "expected_sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337",
+          "actual_sha256": "e90ffd4bb75b001ab41cd956dfbb0a99b574d0b5e8ffc1a64e2887490d658daa"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:scripts/omni/analyze_qwen3_omni_errors.py",
+          "expected_sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337",
+          "actual_sha256": "e90ffd4bb75b001ab41cd956dfbb0a99b574d0b5e8ffc1a64e2887490d658daa"
+        }
+      ]
+    },
     {
       "name": "scripts/audio_ablation_and_raw_upgrade.py",
       "status": "pass",
@@ -1754,21 +1908,21 @@
       "local": {
         "path": "repo:scripts/build_artifact_index.py",
         "exists": true,
-        "bytes": 30785,
-        "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+        "bytes": 32191,
+        "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 30785,
-          "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+          "bytes": 32191,
+          "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
         },
         "hf_model": {
           "path": "hf_model:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 30785,
-          "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+          "bytes": 32191,
+          "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
         }
       },
       "failures": []
@@ -2054,21 +2208,21 @@
       "local": {
         "path": "repo:scripts/validate_mirror_parity.py",
         "exists": true,
-        "bytes": 12642,
-        "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+        "bytes": 13781,
+        "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 12642,
-          "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+          "bytes": 13781,
+          "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
         },
         "hf_model": {
           "path": "hf_model:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 12642,
-          "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+          "bytes": 13781,
+          "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
         }
       },
       "failures": []
@@ -2807,6 +2961,395 @@
       },
       "failures": []
     },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+        "exists": true,
+        "bytes": 3331,
+        "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+        "exists": true,
+        "bytes": 25202,
+        "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "status": "fail",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+        "exists": true,
+        "bytes": 2121,
+        "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2136,
+          "sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2136,
+          "sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2136,
+          "sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+        }
+      },
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "expected_sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd",
+          "actual_sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "expected_sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd",
+          "actual_sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "expected_sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd",
+          "actual_sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+        }
+      ]
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "status": "fail",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+        "exists": true,
+        "bytes": 1320,
+        "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1329,
+          "sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1329,
+          "sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1329,
+          "sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+        }
+      },
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "expected_sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430",
+          "actual_sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "expected_sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430",
+          "actual_sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "expected_sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430",
+          "actual_sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+        }
+      ]
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "status": "fail",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+        "exists": true,
+        "bytes": 572,
+        "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 575,
+          "sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 575,
+          "sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 575,
+          "sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+        }
+      },
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "expected_sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7",
+          "actual_sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "expected_sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7",
+          "actual_sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "expected_sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7",
+          "actual_sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+        }
+      ]
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "status": "fail",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+        "exists": true,
+        "bytes": 408,
+        "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 410,
+          "sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 410,
+          "sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 410,
+          "sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+        }
+      },
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "expected_sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7",
+          "actual_sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "expected_sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7",
+          "actual_sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "expected_sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7",
+          "actual_sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+        }
+      ]
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "status": "fail",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+        "exists": true,
+        "bytes": 1704,
+        "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1715,
+          "sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1715,
+          "sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1715,
+          "sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+        }
+      },
+      "failures": [
+        {
+          "surface": "hf_space",
+          "kind": "hash_mismatch",
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "expected_sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8",
+          "actual_sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+        },
+        {
+          "surface": "hf_artifacts",
+          "kind": "hash_mismatch",
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "expected_sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8",
+          "actual_sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+        },
+        {
+          "surface": "hf_model",
+          "kind": "hash_mismatch",
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "expected_sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8",
+          "actual_sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+        }
+      ]
+    },
+    {
+      "name": "docs/ARTIFACT_GUIDE.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:ARTIFACT_GUIDE.md",
+        "exists": true,
+        "bytes": 16318,
+        "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        },
+        "hf_model": {
+          "path": "hf_model:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "docs/OMNI_MODEL_EXTENSION_CONTRACT.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:OMNI_MODEL_EXTENSION_CONTRACT.md",
+        "exists": true,
+        "bytes": 8900,
+        "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        },
+        "hf_model": {
+          "path": "hf_model:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "docs/QUALITY_GATES.md",
       "status": "pass",
@@ -3061,27 +3604,27 @@
       "local": {
         "path": "repo:PROJECT_STATUS.md",
         "exists": true,
-        "bytes": 8534,
-        "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+        "bytes": 8805,
+        "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         },
         "hf_model": {
           "path": "hf_model:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         }
       },
       "failures": []
@@ -3211,5 +3754,262 @@
       "failures": []
     }
   ],
-  "failures": []
+  "failures": [
+    {
+      "group": "data/artifact_index.json",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:data/artifact_index.json",
+      "expected_sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0",
+      "actual_sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
+    },
+    {
+      "group": "data/artifact_index.json",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:docs/data/artifact_index.json",
+      "expected_sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0",
+      "actual_sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
+    },
+    {
+      "group": "data/artifact_index.json",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:metrics/artifact_index.json",
+      "expected_sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0",
+      "actual_sha256": "2563b854f81b07bfde2880647d0145b511be071a1a274fe1e909ce2be7ce43e1"
+    },
+    {
+      "group": "data/publication_audit.json",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:data/publication_audit.json",
+      "expected_sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d",
+      "actual_sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
+    },
+    {
+      "group": "data/publication_audit.json",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:docs/data/publication_audit.json",
+      "expected_sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d",
+      "actual_sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
+    },
+    {
+      "group": "data/publication_audit.json",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:metrics/publication_audit.json",
+      "expected_sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d",
+      "actual_sha256": "a1741a97c2fb5dee8b9ed8e988b31530128e4fab8b8c458cb8f381e2ad16756c"
+    },
+    {
+      "group": "data/scope_claims_audit.json",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:data/scope_claims_audit.json",
+      "expected_sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3",
+      "actual_sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
+    },
+    {
+      "group": "data/scope_claims_audit.json",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:docs/data/scope_claims_audit.json",
+      "expected_sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3",
+      "actual_sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
+    },
+    {
+      "group": "data/scope_claims_audit.json",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:metrics/scope_claims_audit.json",
+      "expected_sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3",
+      "actual_sha256": "4fb8c088f8ec533b142534b37e9241f8690c8819333434f5d89336c2af8c1c31"
+    },
+    {
+      "group": "data/task_surface_integrity.json",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:data/task_surface_integrity.json",
+      "expected_sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6",
+      "actual_sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
+    },
+    {
+      "group": "data/task_surface_integrity.json",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:docs/data/task_surface_integrity.json",
+      "expected_sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6",
+      "actual_sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
+    },
+    {
+      "group": "data/task_surface_integrity.json",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:metrics/task_surface_integrity.json",
+      "expected_sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6",
+      "actual_sha256": "51c30fe86c558042960e57a252bc6d3c67d95d5a70a8747043a1cdffe57cf53f"
+    },
+    {
+      "group": "data/website_integrity.json",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:data/website_integrity.json",
+      "expected_sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2",
+      "actual_sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
+    },
+    {
+      "group": "data/website_integrity.json",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:docs/data/website_integrity.json",
+      "expected_sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2",
+      "actual_sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
+    },
+    {
+      "group": "data/website_integrity.json",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:metrics/website_integrity.json",
+      "expected_sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2",
+      "actual_sha256": "140d8be179f51351ae55ba7587b7042a1e512e72d9318563a78c96a25e13f830"
+    },
+    {
+      "group": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:scripts/omni/analyze_qwen3_omni_errors.py",
+      "expected_sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337",
+      "actual_sha256": "e90ffd4bb75b001ab41cd956dfbb0a99b574d0b5e8ffc1a64e2887490d658daa"
+    },
+    {
+      "group": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:scripts/omni/analyze_qwen3_omni_errors.py",
+      "expected_sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337",
+      "actual_sha256": "e90ffd4bb75b001ab41cd956dfbb0a99b574d0b5e8ffc1a64e2887490d658daa"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "expected_sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd",
+      "actual_sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "expected_sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd",
+      "actual_sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "expected_sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd",
+      "actual_sha256": "4024fa756edb5a8a9aaac7c213eb411e8d146b109594ad339cc13b08c960bba9"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "expected_sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430",
+      "actual_sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "expected_sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430",
+      "actual_sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "expected_sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430",
+      "actual_sha256": "d995069202708fa456b35aa459ba6e66d90c799b3c4f7b43aa0f6ac4871c986a"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "expected_sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7",
+      "actual_sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "expected_sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7",
+      "actual_sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "expected_sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7",
+      "actual_sha256": "51de0dd0c65d6edc25e78598eebd681fd6ec16ac27de0fa406cc4318023402ad"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "expected_sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7",
+      "actual_sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "expected_sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7",
+      "actual_sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "expected_sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7",
+      "actual_sha256": "1bef2f2a709e2b93a01d1cbb43bb483de5bc5b18b25707c796e1df0bab204171"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "surface": "hf_space",
+      "kind": "hash_mismatch",
+      "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "expected_sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8",
+      "actual_sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "surface": "hf_artifacts",
+      "kind": "hash_mismatch",
+      "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "expected_sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8",
+      "actual_sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+    },
+    {
+      "group": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "surface": "hf_model",
+      "kind": "hash_mismatch",
+      "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "expected_sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8",
+      "actual_sha256": "5c0e94caf4fe1eb26565e0bb796cd3c1eed2741b98a528f38317bcfb6a4c2e23"
+    }
+  ]
 }

data/omni_finetune_verified_result.json

@@ -67,7 +67,28 @@
     "audit_status": "pass",
     "contains_raw_xperience10m_data": false,
     "contains_qwen_base_weights": false,
-    "contains_lora_weights": false
+    "contains_lora_weights": false,
+    "error_analysis": {
+      "status": "pass",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "markdown_report": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "groupings": [
+        "episode",
+        "action_family",
+        "train_seen_status",
+        "required_modality_state",
+        "object_category"
+      ],
+      "key_readouts": {
+        "parsed_prediction_rate": 0.8772321428571429,
+        "weakest_action_family": "locomotion",
+        "weakest_action_family_samples": 23,
+        "weakest_action_family_parsed_prediction_rate": 0.2608695652173913,
+        "seen_action_exact_rate": 0.04580152671755725,
+        "unseen_action_exact_rate": 0.015772870662460567,
+        "required_modality_state": "rrd_missing_only_required_modalities_present"
+      }
+    }
   },
   "required_next_steps": [
     "Improve JSON-format reliability through prompt, decoding, constrained parsing, or target formatting changes.",

data/project_status.json

@@ -180,10 +180,12 @@
       "evidence": [
         "docs/data/omni_finetune_verified_result.json",
         "results/omni_finetune/verified_public/",
+        "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/",
         "scripts/omni/package_verified_omni_result.py",
-        "scripts/omni/audit_verified_omni_package.py"
+        "scripts/omni/audit_verified_omni_package.py",
+        "scripts/omni/analyze_qwen3_omni_errors.py"
       ],
-      "readout": "The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, and 448 test predictions. JSON validity is 87.50%, below the 98% target, so it is a stronger diagnostic baseline but not a strong model-quality result."
+      "readout": "The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, 448 test predictions, and derived error-analysis tables by episode, action family, train-seen status, required-modality state, and object category. JSON validity is 87.50%, below the 98% target, so it is a diagnostic baseline but not a strong model-quality result."
     },
     {
       "area": "Raw Xperience-10M redistribution",

data/publication_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:38:05+00:00",
+  "generated_at_utc": "2026-06-06T14:54:02+00:00",
   "checks": [
     {
       "name": "required_publication_assets_present",
@@ -182,8 +182,8 @@
     "github_repo": {
       "root": "repo",
       "exists": true,
-      "file_count": 442,
-      "text_file_count": 372,
+      "file_count": 450,
+      "text_file_count": 380,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -193,8 +193,8 @@
     "hf_space_bundle": {
       "root": "hf_publish/space",
       "exists": true,
-      "file_count": 356,
-      "text_file_count": 286,
+      "file_count": 363,
+      "text_file_count": 293,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -204,8 +204,8 @@
     "hf_artifact_bundle": {
       "root": "hf_publish/artifacts",
       "exists": true,
-      "file_count": 514,
-      "text_file_count": 420,
+      "file_count": 522,
+      "text_file_count": 428,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -215,8 +215,8 @@
     "hf_model_bundle": {
       "root": "hf_publish/model",
       "exists": true,
-      "file_count": 701,
-      "text_file_count": 572,
+      "file_count": 709,
+      "text_file_count": 580,
       "largest_file": {
         "path": "pytorch_model.bin",
         "bytes": 93495480

data/scope_claims_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:35:59+00:00",
+  "generated_at_utc": "2026-06-06T14:54:01+00:00",
   "summary": {
     "qwen3_omni_verified_diagnostic_pilot": true,
     "dataset_manifest_num_episodes": 119,

data/task_surface_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:35:59+00:00",
+  "generated_at_utc": "2026-06-06T14:53:59+00:00",
   "summary": {
     "task_count": 12,
     "expected_task_count": 12,
@@ -64,15 +64,21 @@
       "observed": "timeline_action"
     },
     {
-      "name": "timeline_action: public_field_input_short_is_human_readable",
+      "name": "timeline_action: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "20-frame multimodal window",
+      "value": "Recognize the current manipulation action from synchronized visual, motion, inertial, pose, and annotation context.",
       "raw_hits": []
     },
     {
-      "name": "timeline_action: public_field_card_blurb_is_human_readable",
+      "name": "timeline_action: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Recognize the current manipulation action from synchronized visual, motion, inertial, pose, and annotation context.",
+      "value": "Egocentric Action Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "timeline_action: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "20-frame multimodal window",
       "raw_hits": []
     },
     {
@@ -88,9 +94,9 @@
       "raw_hits": []
     },
     {
-      "name": "timeline_action: public_field_research_name_is_human_readable",
+      "name": "timeline_action: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Egocentric Action Recognition",
+      "value": "window features -> action label builder -> classifier",
       "raw_hits": []
     },
     {
@@ -99,12 +105,6 @@
       "value": "Look at one short multimodal window and name what action is happening now.",
       "raw_hits": []
     },
-    {
-      "name": "timeline_action: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "window features -> action label builder -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "timeline_action: known_task_family",
       "status": "pass",
@@ -184,15 +184,21 @@
       "observed": "timeline_subtask"
     },
     {
-      "name": "timeline_subtask: public_field_input_short_is_human_readable",
+      "name": "timeline_subtask: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "20-frame multimodal window",
+      "value": "Recognize the broader activity stage so fine actions become a readable procedure timeline.",
       "raw_hits": []
     },
     {
-      "name": "timeline_subtask: public_field_card_blurb_is_human_readable",
+      "name": "timeline_subtask: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Recognize the broader activity stage so fine actions become a readable procedure timeline.",
+      "value": "Temporal Subtask Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "timeline_subtask: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "20-frame multimodal window",
       "raw_hits": []
     },
     {
@@ -208,9 +214,9 @@
       "raw_hits": []
     },
     {
-      "name": "timeline_subtask: public_field_research_name_is_human_readable",
+      "name": "timeline_subtask: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Temporal Subtask Recognition",
+      "value": "window features -> subtask label builder -> classifier",
       "raw_hits": []
     },
     {
@@ -219,12 +225,6 @@
       "value": "Predict the higher-level task stage for the current window.",
       "raw_hits": []
     },
-    {
-      "name": "timeline_subtask: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "window features -> subtask label builder -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "timeline_subtask: known_task_family",
       "status": "pass",
@@ -304,15 +304,21 @@
       "observed": "transition_detection"
     },
     {
-      "name": "transition_detection: public_field_input_short_is_human_readable",
+      "name": "transition_detection: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current window with boundary target",
+      "value": "Detect the local moment where the episode changes from one action segment to the next.",
       "raw_hits": []
     },
     {
-      "name": "transition_detection: public_field_card_blurb_is_human_readable",
+      "name": "transition_detection: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Detect the local moment where the episode changes from one action segment to the next.",
+      "value": "Temporal Action Segmentation",
+      "raw_hits": []
+    },
+    {
+      "name": "transition_detection: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current window with boundary target",
       "raw_hits": []
     },
     {
@@ -328,9 +334,9 @@
       "raw_hits": []
     },
     {
-      "name": "transition_detection: public_field_research_name_is_human_readable",
+      "name": "transition_detection: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Temporal Action Segmentation",
+      "value": "action changes -> boundary labels -> binary classifier",
       "raw_hits": []
     },
     {
@@ -339,12 +345,6 @@
       "value": "Detect whether the current window is near a boundary between actions.",
       "raw_hits": []
     },
-    {
-      "name": "transition_detection: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "action changes -> boundary labels -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "transition_detection: known_task_family",
       "status": "pass",
@@ -422,15 +422,21 @@
       "observed": "next_action"
     },
     {
-      "name": "next_action: public_field_input_short_is_human_readable",
+      "name": "next_action: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current window at time t",
+      "value": "Forecast the near-future action from the current observations only.",
       "raw_hits": []
     },
     {
-      "name": "next_action: public_field_card_blurb_is_human_readable",
+      "name": "next_action: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Forecast the near-future action from the current observations only.",
+      "value": "Short-Horizon Intention Prediction",
+      "raw_hits": []
+    },
+    {
+      "name": "next_action: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current window at time t",
       "raw_hits": []
     },
     {
@@ -446,9 +452,9 @@
       "raw_hits": []
     },
     {
-      "name": "next_action: public_field_research_name_is_human_readable",
+      "name": "next_action: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Short-Horizon Intention Prediction",
+      "value": "current features -> future label shift -> classifier",
       "raw_hits": []
     },
     {
@@ -457,12 +463,6 @@
       "value": "Use the current window to guess the action that will happen shortly after it.",
       "raw_hits": []
     },
-    {
-      "name": "next_action: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "current features -> future label shift -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "next_action: known_task_family",
       "status": "pass",
@@ -540,15 +540,21 @@
       "observed": "hand_trajectory_forecast"
     },
     {
-      "name": "hand_trajectory_forecast: public_field_input_short_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current multimodal window",
+      "value": "Predict the future 3D left/right hand path from the current multimodal state.",
       "raw_hits": []
     },
     {
-      "name": "hand_trajectory_forecast: public_field_card_blurb_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict the future 3D left/right hand path from the current multimodal state.",
+      "value": "3D Hand Motion Forecasting",
+      "raw_hits": []
+    },
+    {
+      "name": "hand_trajectory_forecast: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current multimodal window",
       "raw_hits": []
     },
     {
@@ -564,9 +570,9 @@
       "raw_hits": []
     },
     {
-      "name": "hand_trajectory_forecast: public_field_research_name_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "3D Hand Motion Forecasting",
+      "value": "current features -> future mocap target -> regression head",
       "raw_hits": []
     },
     {
@@ -575,12 +581,6 @@
       "value": "Predict where the hands will move over the next few frames.",
       "raw_hits": []
     },
-    {
-      "name": "hand_trajectory_forecast: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "current features -> future mocap target -> regression head",
-      "raw_hits": []
-    },
     {
       "name": "hand_trajectory_forecast: known_task_family",
       "status": "pass",
@@ -658,15 +658,21 @@
       "observed": "contact_prediction"
     },
     {
-      "name": "contact_prediction: public_field_input_short_is_human_readable",
+      "name": "contact_prediction: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "non-contact, non-caption features",
+      "value": "Predict whether body or hand contact with the scene is occurring without leaking contact labels.",
       "raw_hits": []
     },
     {
-      "name": "contact_prediction: public_field_card_blurb_is_human_readable",
+      "name": "contact_prediction: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict whether body or hand contact with the scene is occurring without leaking contact labels.",
+      "value": "Human-Object Contact Prediction",
+      "raw_hits": []
+    },
+    {
+      "name": "contact_prediction: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "non-contact, non-caption features",
       "raw_hits": []
     },
     {
@@ -682,9 +688,9 @@
       "raw_hits": []
     },
     {
-      "name": "contact_prediction: public_field_research_name_is_human_readable",
+      "name": "contact_prediction: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Human-Object Contact Prediction",
+      "value": "feature filter -> contact target -> binary classifier",
       "raw_hits": []
     },
     {
@@ -693,12 +699,6 @@
       "value": "Predict whether the body or hand is in contact with something.",
       "raw_hits": []
     },
-    {
-      "name": "contact_prediction: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "feature filter -> contact target -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "contact_prediction: known_task_family",
       "status": "pass",
@@ -774,15 +774,21 @@
       "observed": "object_relevance"
     },
     {
-      "name": "object_relevance: public_field_input_short_is_human_readable",
+      "name": "object_relevance: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "non-caption multimodal features",
+      "value": "Infer which objects are relevant to the current manipulation window from non-caption features.",
       "raw_hits": []
     },
     {
-      "name": "object_relevance: public_field_card_blurb_is_human_readable",
+      "name": "object_relevance: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Infer which objects are relevant to the current manipulation window from non-caption features.",
+      "value": "Object-Centric Interaction Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "object_relevance: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "non-caption multimodal features",
       "raw_hits": []
     },
     {
@@ -798,9 +804,9 @@
       "raw_hits": []
     },
     {
-      "name": "object_relevance: public_field_research_name_is_human_readable",
+      "name": "object_relevance: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Object-Centric Interaction Recognition",
+      "value": "object vocabulary -> multi-hot labels -> sigmoid heads",
       "raw_hits": []
     },
     {
@@ -809,12 +815,6 @@
       "value": "Predict which objects matter in the current window.",
       "raw_hits": []
     },
-    {
-      "name": "object_relevance: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "object vocabulary -> multi-hot labels -> sigmoid heads",
-      "raw_hits": []
-    },
     {
       "name": "object_relevance: known_task_family",
       "status": "pass",
@@ -892,15 +892,21 @@
       "observed": "caption_grounding"
     },
     {
-      "name": "caption_grounding: public_field_input_short_is_human_readable",
+      "name": "caption_grounding: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "text-like query and candidate windows",
+      "value": "Retrieve the matching time window for an annotation-derived text query.",
       "raw_hits": []
     },
     {
-      "name": "caption_grounding: public_field_card_blurb_is_human_readable",
+      "name": "caption_grounding: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Retrieve the matching time window for an annotation-derived text query.",
+      "value": "Language-to-Moment Grounding",
+      "raw_hits": []
+    },
+    {
+      "name": "caption_grounding: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "text-like query and candidate windows",
       "raw_hits": []
     },
     {
@@ -916,9 +922,9 @@
       "raw_hits": []
     },
     {
-      "name": "caption_grounding: public_field_research_name_is_human_readable",
+      "name": "caption_grounding: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Language-to-Moment Grounding",
+      "value": "query features -> candidate index -> cosine ranker",
       "raw_hits": []
     },
     {
@@ -927,12 +933,6 @@
       "value": "Given a text-like query from annotation, find the matching time window.",
       "raw_hits": []
     },
-    {
-      "name": "caption_grounding: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "query features -> candidate index -> cosine ranker",
-      "raw_hits": []
-    },
     {
       "name": "caption_grounding: known_task_family",
       "status": "pass",
@@ -1008,15 +1008,21 @@
       "observed": "cross_modal_retrieval"
     },
     {
-      "name": "cross_modal_retrieval: public_field_input_short_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion/IMU/pose query; depth/video candidates",
+      "value": "Use motion, IMU, and camera-pose signals to retrieve the matching depth/video window.",
       "raw_hits": []
     },
     {
-      "name": "cross_modal_retrieval: public_field_card_blurb_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Use motion, IMU, and camera-pose signals to retrieve the matching depth/video window.",
+      "value": "Multimodal Representation Retrieval",
+      "raw_hits": []
+    },
+    {
+      "name": "cross_modal_retrieval: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion/IMU/pose query; depth/video candidates",
       "raw_hits": []
     },
     {
@@ -1032,9 +1038,9 @@
       "raw_hits": []
     },
     {
-      "name": "cross_modal_retrieval: public_field_research_name_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Multimodal Representation Retrieval",
+      "value": "modality split -> projection -> nearest-neighbor ranker",
       "raw_hits": []
     },
     {
@@ -1043,12 +1049,6 @@
       "value": "Use one group of modalities to retrieve the matching window from another group.",
       "raw_hits": []
     },
-    {
-      "name": "cross_modal_retrieval: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "modality split -> projection -> nearest-neighbor ranker",
-      "raw_hits": []
-    },
     {
       "name": "cross_modal_retrieval: known_task_family",
       "status": "pass",
@@ -1126,15 +1126,21 @@
       "observed": "modality_reconstruction"
     },
     {
-      "name": "modality_reconstruction: public_field_input_short_is_human_readable",
+      "name": "modality_reconstruction: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion, IMU, and camera/pose features",
+      "value": "Predict compressed depth/video feature vectors from motion, IMU, and camera-pose features.",
       "raw_hits": []
     },
     {
-      "name": "modality_reconstruction: public_field_card_blurb_is_human_readable",
+      "name": "modality_reconstruction: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict compressed depth/video feature vectors from motion, IMU, and camera-pose features.",
+      "value": "Modality Feature Reconstruction",
+      "raw_hits": []
+    },
+    {
+      "name": "modality_reconstruction: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion, IMU, and camera/pose features",
       "raw_hits": []
     },
     {
@@ -1150,9 +1156,9 @@
       "raw_hits": []
     },
     {
-      "name": "modality_reconstruction: public_field_research_name_is_human_readable",
+      "name": "modality_reconstruction: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Modality Feature Reconstruction",
+      "value": "source-target split -> scaler -> regression head",
       "raw_hits": []
     },
     {
@@ -1161,12 +1167,6 @@
       "value": "Predict one modality feature block from other modality blocks.",
       "raw_hits": []
     },
-    {
-      "name": "modality_reconstruction: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "source-target split -> scaler -> regression head",
-      "raw_hits": []
-    },
     {
       "name": "modality_reconstruction: known_task_family",
       "status": "pass",
@@ -1243,12 +1243,6 @@
       "status": "pass",
       "observed": "temporal_order"
     },
-    {
-      "name": "temporal_order: public_field_input_short_is_human_readable",
-      "status": "pass",
-      "value": "two adjacent windows plus difference vector",
-      "raw_hits": []
-    },
     {
       "name": "temporal_order: public_field_card_blurb_is_human_readable",
       "status": "pass",
@@ -1256,27 +1250,27 @@
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_display_name_is_human_readable",
+      "name": "temporal_order: public_field_research_name_is_human_readable",
       "status": "pass",
       "value": "Temporal Order Verification",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_output_short_is_human_readable",
+      "name": "temporal_order: public_field_input_short_is_human_readable",
       "status": "pass",
-      "value": "correct or reversed",
+      "value": "two adjacent windows plus difference vector",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_research_name_is_human_readable",
+      "name": "temporal_order: public_field_display_name_is_human_readable",
       "status": "pass",
       "value": "Temporal Order Verification",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_plain_goal_is_human_readable",
+      "name": "temporal_order: public_field_output_short_is_human_readable",
       "status": "pass",
-      "value": "Tell whether two nearby windows are in the correct time order.",
+      "value": "correct or reversed",
       "raw_hits": []
     },
     {
@@ -1285,6 +1279,12 @@
       "value": "pair builder -> feature combiner -> binary classifier",
       "raw_hits": []
     },
+    {
+      "name": "temporal_order: public_field_plain_goal_is_human_readable",
+      "status": "pass",
+      "value": "Tell whether two nearby windows are in the correct time order.",
+      "raw_hits": []
+    },
     {
       "name": "temporal_order: known_task_family",
       "status": "pass",
@@ -1360,15 +1360,21 @@
       "observed": "misalignment_detection"
     },
     {
-      "name": "misalignment_detection: public_field_input_short_is_human_readable",
+      "name": "misalignment_detection: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion-side and visual/depth-side feature groups",
+      "value": "Detect whether motion and visual/depth streams have been artificially shifted out of sync.",
       "raw_hits": []
     },
     {
-      "name": "misalignment_detection: public_field_card_blurb_is_human_readable",
+      "name": "misalignment_detection: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Detect whether motion and visual/depth streams have been artificially shifted out of sync.",
+      "value": "Cross-Modal Misalignment Detection",
+      "raw_hits": []
+    },
+    {
+      "name": "misalignment_detection: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion-side and visual/depth-side feature groups",
       "raw_hits": []
     },
     {
@@ -1384,9 +1390,9 @@
       "raw_hits": []
     },
     {
-      "name": "misalignment_detection: public_field_research_name_is_human_readable",
+      "name": "misalignment_detection: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Cross-Modal Misalignment Detection",
+      "value": "aligned/shifted pairs -> feature combiner -> binary classifier",
       "raw_hits": []
     },
     {
@@ -1395,12 +1401,6 @@
       "value": "Detect when modalities that should match are shifted out of sync.",
       "raw_hits": []
     },
-    {
-      "name": "misalignment_detection: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "aligned/shifted pairs -> feature combiner -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "misalignment_detection: known_task_family",
       "status": "pass",

data/website_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:36:10+00:00",
+  "generated_at_utc": "2026-06-06T14:54:01+00:00",
   "docs_root": "docs",
   "site_base": "/ropedia-xperience-10m-task-suite/",
   "summary": {
@@ -251,7 +251,7 @@
     },
     {
       "path": "data/artifact_index.json",
-      "bytes": 37736,
+      "bytes": 39486,
       "top_level_type": "dict"
     },
     {
@@ -291,7 +291,7 @@
     },
     {
       "path": "data/mirror_parity.json",
-      "bytes": 111950,
+      "bytes": 126335,
       "top_level_type": "dict"
     },
     {
@@ -301,7 +301,7 @@
     },
     {
       "path": "data/omni_finetune_verified_result.json",
-      "bytes": 3145,
+      "bytes": 4142,
       "top_level_type": "dict"
     },
     {
@@ -321,7 +321,7 @@
     },
     {
       "path": "data/project_status.json",
-      "bytes": 10977,
+      "bytes": 11274,
       "top_level_type": "dict"
     },
     {

docs/data/artifact_index.json

@@ -1,12 +1,12 @@
 {
   "title": "Ropedia Xperience-10M Task Suite Artifact Index",
-  "generated_at_utc": "2026-06-06T14:35:42+00:00",
+  "generated_at_utc": "2026-06-06T14:53:45+00:00",
   "status": "pass",
-  "artifact_count": 83,
+  "artifact_count": 86,
   "missing": [],
   "by_kind": {
     "project_path": 14,
-    "scaleup_contract": 6,
+    "scaleup_contract": 7,
     "project_scope": 1,
     "source_alignment": 5,
     "publication_workflow": 3,
@@ -28,7 +28,7 @@
     "onboarding_doc": 1,
     "generated_figure": 3,
     "generated_figure_assets": 1,
-    "scaleup_status": 2,
+    "scaleup_status": 4,
     "citation": 1,
     "license": 1
   },
@@ -63,8 +63,8 @@
       "surface": "repo_hf",
       "shows": "Gives a compact current-state table for first-pass readers.",
       "exists": true,
-      "bytes": 8534,
-      "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+      "bytes": 8805,
+      "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
     },
     {
       "id": "project_status_json",
@@ -74,8 +74,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable copy of the current project status for website and HF mirrors.",
       "exists": true,
-      "bytes": 10977,
-      "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+      "bytes": 11274,
+      "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
     },
     {
       "id": "research_roadmap",
@@ -187,6 +187,17 @@
       "bytes": 6519,
       "sha256": "a3773fc681e298325e2be80556d6be6e7e30b90ba22ee24b66633f07ff9c4ea4"
     },
+    {
+      "id": "qwen3_omni_error_analysis_script",
+      "title": "Qwen3-Omni held-out error-analysis script",
+      "path": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "kind": "scaleup_contract",
+      "surface": "repo_hf",
+      "shows": "Computes public-safe held-out error-analysis tables by episode, action family, train-seen status, required-modality state, and object category.",
+      "exists": true,
+      "bytes": 15676,
+      "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+    },
     {
       "id": "additional_development_directions",
       "title": "Additional development directions",
@@ -250,8 +261,8 @@
       "surface": "repo_hf",
       "shows": "Gives the human-readable map from project scope to data, tasks, platform mirrors, and scale-up status.",
       "exists": true,
-      "bytes": 15660,
-      "sha256": "a9ad335b82c35a5ac102428663ffae1c8798e90e45cc5e795c3a499b4563b417"
+      "bytes": 16318,
+      "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
     },
     {
       "id": "official_dataset_card_alignment",
@@ -695,8 +706,8 @@
       "surface": "repo_hf",
       "shows": "Generates the selective artifact catalog from local files.",
       "exists": true,
-      "bytes": 30785,
-      "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+      "bytes": 32191,
+      "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
     },
     {
       "id": "publication_audit",
@@ -731,7 +742,7 @@
       "volatile": true,
       "shows": "Confirms prepared GitHub/HF Space/artifact/model mirrors share the same critical data, figure, website HTML, and validator files.",
       "exists": true,
-      "bytes": 111950,
+      "bytes": 126335,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -933,6 +944,28 @@
       "bytes": 3076,
       "sha256": "23b87581cfc1d95b0af118a0dbb4e601f42fc6bad608759490e13a9a1ef73205"
     },
+    {
+      "id": "qwen3_omni_error_analysis_report",
+      "title": "Qwen3-Omni held-out error-analysis report",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "kind": "scaleup_status",
+      "surface": "repo_hf",
+      "shows": "Summarizes validation-aware Qwen3-Omni held-out failures by episode, action family, train-seen status, required-modality state, and object category.",
+      "exists": true,
+      "bytes": 3331,
+      "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+    },
+    {
+      "id": "qwen3_omni_error_analysis_json",
+      "title": "Qwen3-Omni held-out error-analysis JSON",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "kind": "scaleup_status",
+      "surface": "repo_hf",
+      "shows": "Machine-readable Qwen3-Omni held-out error analysis with grouped metrics and sanitized failure examples.",
+      "exists": true,
+      "bytes": 25202,
+      "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+    },
     {
       "id": "citation",
       "title": "Citation metadata",

docs/data/mirror_parity.json

@@ -1,9 +1,9 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:37:36+00:00",
+  "generated_at_utc": "2026-06-06T14:56:44+00:00",
   "hf_root": "hf_publish",
   "summary": {
-    "group_count": 104,
+    "group_count": 114,
     "failure_count": 0,
     "failures_by_surface": {}
   },
@@ -102,27 +102,27 @@
       "local": {
         "path": "repo:docs/data/artifact_index.json",
         "exists": true,
-        "bytes": 37736,
-        "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+        "bytes": 39486,
+        "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
         },
         "hf_model": {
           "path": "hf_model:metrics/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
         }
       },
       "failures": []
@@ -350,27 +350,27 @@
       "local": {
         "path": "repo:docs/data/omni_finetune_verified_result.json",
         "exists": true,
-        "bytes": 3145,
-        "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+        "bytes": 4142,
+        "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         },
         "hf_model": {
           "path": "hf_model:metrics/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         }
       },
       "failures": []
@@ -474,27 +474,27 @@
       "local": {
         "path": "repo:docs/data/project_status.json",
         "exists": true,
-        "bytes": 10977,
-        "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+        "bytes": 11274,
+        "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         },
         "hf_model": {
           "path": "hf_model:metrics/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         }
       },
       "failures": []
@@ -506,26 +506,26 @@
         "path": "repo:docs/data/publication_audit.json",
         "exists": true,
         "bytes": 7237,
-        "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+        "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
         },
         "hf_model": {
           "path": "hf_model:metrics/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
         }
       },
       "failures": []
@@ -816,26 +816,26 @@
         "path": "repo:docs/data/scope_claims_audit.json",
         "exists": true,
         "bytes": 20823,
-        "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+        "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
         },
         "hf_model": {
           "path": "hf_model:metrics/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
         }
       },
       "failures": []
@@ -940,26 +940,26 @@
         "path": "repo:docs/data/task_surface_integrity.json",
         "exists": true,
         "bytes": 45779,
-        "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+        "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
         },
         "hf_model": {
           "path": "hf_model:metrics/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
         }
       },
       "failures": []
@@ -1002,26 +1002,26 @@
         "path": "repo:docs/data/website_integrity.json",
         "exists": true,
         "bytes": 15221,
-        "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+        "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
         },
         "hf_model": {
           "path": "hf_model:metrics/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
         }
       },
       "failures": []
@@ -1723,6 +1723,31 @@
       },
       "failures": []
     },
+    {
+      "name": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "status": "pass",
+      "local": {
+        "path": "repo:scripts/omni/analyze_qwen3_omni_errors.py",
+        "exists": true,
+        "bytes": 15676,
+        "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+      },
+      "mirrors": {
+        "hf_artifacts": {
+          "path": "hf_artifacts:scripts/omni/analyze_qwen3_omni_errors.py",
+          "exists": true,
+          "bytes": 15676,
+          "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+        },
+        "hf_model": {
+          "path": "hf_model:scripts/omni/analyze_qwen3_omni_errors.py",
+          "exists": true,
+          "bytes": 15676,
+          "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "scripts/audio_ablation_and_raw_upgrade.py",
       "status": "pass",
@@ -1754,21 +1779,21 @@
       "local": {
         "path": "repo:scripts/build_artifact_index.py",
         "exists": true,
-        "bytes": 30785,
-        "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+        "bytes": 32191,
+        "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 30785,
-          "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+          "bytes": 32191,
+          "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
         },
         "hf_model": {
           "path": "hf_model:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 30785,
-          "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+          "bytes": 32191,
+          "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
         }
       },
       "failures": []
@@ -2054,21 +2079,21 @@
       "local": {
         "path": "repo:scripts/validate_mirror_parity.py",
         "exists": true,
-        "bytes": 12642,
-        "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+        "bytes": 13781,
+        "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 12642,
-          "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+          "bytes": 13781,
+          "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
         },
         "hf_model": {
           "path": "hf_model:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 12642,
-          "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+          "bytes": 13781,
+          "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
         }
       },
       "failures": []
@@ -2807,6 +2832,285 @@
       },
       "failures": []
     },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+        "exists": true,
+        "bytes": 3331,
+        "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+        "exists": true,
+        "bytes": 25202,
+        "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+        "exists": true,
+        "bytes": 2121,
+        "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2121,
+          "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2121,
+          "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2121,
+          "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+        "exists": true,
+        "bytes": 1320,
+        "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1320,
+          "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1320,
+          "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1320,
+          "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+        "exists": true,
+        "bytes": 572,
+        "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 572,
+          "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 572,
+          "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 572,
+          "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+        "exists": true,
+        "bytes": 408,
+        "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 408,
+          "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 408,
+          "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 408,
+          "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+        "exists": true,
+        "bytes": 1704,
+        "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1704,
+          "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1704,
+          "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1704,
+          "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "docs/ARTIFACT_GUIDE.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:ARTIFACT_GUIDE.md",
+        "exists": true,
+        "bytes": 16318,
+        "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        },
+        "hf_model": {
+          "path": "hf_model:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "docs/OMNI_MODEL_EXTENSION_CONTRACT.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:OMNI_MODEL_EXTENSION_CONTRACT.md",
+        "exists": true,
+        "bytes": 8900,
+        "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        },
+        "hf_model": {
+          "path": "hf_model:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "docs/QUALITY_GATES.md",
       "status": "pass",
@@ -3061,27 +3365,27 @@
       "local": {
         "path": "repo:PROJECT_STATUS.md",
         "exists": true,
-        "bytes": 8534,
-        "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+        "bytes": 8805,
+        "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         },
         "hf_model": {
           "path": "hf_model:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         }
       },
       "failures": []

docs/data/omni_finetune_verified_result.json

@@ -67,7 +67,28 @@
     "audit_status": "pass",
     "contains_raw_xperience10m_data": false,
     "contains_qwen_base_weights": false,
-    "contains_lora_weights": false
+    "contains_lora_weights": false,
+    "error_analysis": {
+      "status": "pass",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "markdown_report": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "groupings": [
+        "episode",
+        "action_family",
+        "train_seen_status",
+        "required_modality_state",
+        "object_category"
+      ],
+      "key_readouts": {
+        "parsed_prediction_rate": 0.8772321428571429,
+        "weakest_action_family": "locomotion",
+        "weakest_action_family_samples": 23,
+        "weakest_action_family_parsed_prediction_rate": 0.2608695652173913,
+        "seen_action_exact_rate": 0.04580152671755725,
+        "unseen_action_exact_rate": 0.015772870662460567,
+        "required_modality_state": "rrd_missing_only_required_modalities_present"
+      }
+    }
   },
   "required_next_steps": [
     "Improve JSON-format reliability through prompt, decoding, constrained parsing, or target formatting changes.",

docs/data/project_status.json

@@ -180,10 +180,12 @@
       "evidence": [
         "docs/data/omni_finetune_verified_result.json",
         "results/omni_finetune/verified_public/",
+        "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/",
         "scripts/omni/package_verified_omni_result.py",
-        "scripts/omni/audit_verified_omni_package.py"
+        "scripts/omni/audit_verified_omni_package.py",
+        "scripts/omni/analyze_qwen3_omni_errors.py"
       ],
-      "readout": "The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, and 448 test predictions. JSON validity is 87.50%, below the 98% target, so it is a stronger diagnostic baseline but not a strong model-quality result."
+      "readout": "The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, 448 test predictions, and derived error-analysis tables by episode, action family, train-seen status, required-modality state, and object category. JSON validity is 87.50%, below the 98% target, so it is a diagnostic baseline but not a strong model-quality result."
     },
     {
       "area": "Raw Xperience-10M redistribution",

docs/data/publication_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:38:05+00:00",
+  "generated_at_utc": "2026-06-06T14:54:02+00:00",
   "checks": [
     {
       "name": "required_publication_assets_present",
@@ -182,8 +182,8 @@
     "github_repo": {
       "root": "repo",
       "exists": true,
-      "file_count": 442,
-      "text_file_count": 372,
+      "file_count": 450,
+      "text_file_count": 380,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -193,8 +193,8 @@
     "hf_space_bundle": {
       "root": "hf_publish/space",
       "exists": true,
-      "file_count": 356,
-      "text_file_count": 286,
+      "file_count": 363,
+      "text_file_count": 293,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -204,8 +204,8 @@
     "hf_artifact_bundle": {
       "root": "hf_publish/artifacts",
       "exists": true,
-      "file_count": 514,
-      "text_file_count": 420,
+      "file_count": 522,
+      "text_file_count": 428,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -215,8 +215,8 @@
     "hf_model_bundle": {
       "root": "hf_publish/model",
       "exists": true,
-      "file_count": 701,
-      "text_file_count": 572,
+      "file_count": 709,
+      "text_file_count": 580,
       "largest_file": {
         "path": "pytorch_model.bin",
         "bytes": 93495480

docs/data/scope_claims_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:35:59+00:00",
+  "generated_at_utc": "2026-06-06T14:54:01+00:00",
   "summary": {
     "qwen3_omni_verified_diagnostic_pilot": true,
     "dataset_manifest_num_episodes": 119,

docs/data/task_surface_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:35:59+00:00",
+  "generated_at_utc": "2026-06-06T14:53:59+00:00",
   "summary": {
     "task_count": 12,
     "expected_task_count": 12,
@@ -64,15 +64,21 @@
       "observed": "timeline_action"
     },
     {
-      "name": "timeline_action: public_field_input_short_is_human_readable",
+      "name": "timeline_action: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "20-frame multimodal window",
+      "value": "Recognize the current manipulation action from synchronized visual, motion, inertial, pose, and annotation context.",
       "raw_hits": []
     },
     {
-      "name": "timeline_action: public_field_card_blurb_is_human_readable",
+      "name": "timeline_action: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Recognize the current manipulation action from synchronized visual, motion, inertial, pose, and annotation context.",
+      "value": "Egocentric Action Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "timeline_action: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "20-frame multimodal window",
       "raw_hits": []
     },
     {
@@ -88,9 +94,9 @@
       "raw_hits": []
     },
     {
-      "name": "timeline_action: public_field_research_name_is_human_readable",
+      "name": "timeline_action: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Egocentric Action Recognition",
+      "value": "window features -> action label builder -> classifier",
       "raw_hits": []
     },
     {
@@ -99,12 +105,6 @@
       "value": "Look at one short multimodal window and name what action is happening now.",
       "raw_hits": []
     },
-    {
-      "name": "timeline_action: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "window features -> action label builder -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "timeline_action: known_task_family",
       "status": "pass",
@@ -184,15 +184,21 @@
       "observed": "timeline_subtask"
     },
     {
-      "name": "timeline_subtask: public_field_input_short_is_human_readable",
+      "name": "timeline_subtask: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "20-frame multimodal window",
+      "value": "Recognize the broader activity stage so fine actions become a readable procedure timeline.",
       "raw_hits": []
     },
     {
-      "name": "timeline_subtask: public_field_card_blurb_is_human_readable",
+      "name": "timeline_subtask: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Recognize the broader activity stage so fine actions become a readable procedure timeline.",
+      "value": "Temporal Subtask Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "timeline_subtask: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "20-frame multimodal window",
       "raw_hits": []
     },
     {
@@ -208,9 +214,9 @@
       "raw_hits": []
     },
     {
-      "name": "timeline_subtask: public_field_research_name_is_human_readable",
+      "name": "timeline_subtask: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Temporal Subtask Recognition",
+      "value": "window features -> subtask label builder -> classifier",
       "raw_hits": []
     },
     {
@@ -219,12 +225,6 @@
       "value": "Predict the higher-level task stage for the current window.",
       "raw_hits": []
     },
-    {
-      "name": "timeline_subtask: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "window features -> subtask label builder -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "timeline_subtask: known_task_family",
       "status": "pass",
@@ -304,15 +304,21 @@
       "observed": "transition_detection"
     },
     {
-      "name": "transition_detection: public_field_input_short_is_human_readable",
+      "name": "transition_detection: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current window with boundary target",
+      "value": "Detect the local moment where the episode changes from one action segment to the next.",
       "raw_hits": []
     },
     {
-      "name": "transition_detection: public_field_card_blurb_is_human_readable",
+      "name": "transition_detection: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Detect the local moment where the episode changes from one action segment to the next.",
+      "value": "Temporal Action Segmentation",
+      "raw_hits": []
+    },
+    {
+      "name": "transition_detection: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current window with boundary target",
       "raw_hits": []
     },
     {
@@ -328,9 +334,9 @@
       "raw_hits": []
     },
     {
-      "name": "transition_detection: public_field_research_name_is_human_readable",
+      "name": "transition_detection: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Temporal Action Segmentation",
+      "value": "action changes -> boundary labels -> binary classifier",
       "raw_hits": []
     },
     {
@@ -339,12 +345,6 @@
       "value": "Detect whether the current window is near a boundary between actions.",
       "raw_hits": []
     },
-    {
-      "name": "transition_detection: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "action changes -> boundary labels -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "transition_detection: known_task_family",
       "status": "pass",
@@ -422,15 +422,21 @@
       "observed": "next_action"
     },
     {
-      "name": "next_action: public_field_input_short_is_human_readable",
+      "name": "next_action: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current window at time t",
+      "value": "Forecast the near-future action from the current observations only.",
       "raw_hits": []
     },
     {
-      "name": "next_action: public_field_card_blurb_is_human_readable",
+      "name": "next_action: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Forecast the near-future action from the current observations only.",
+      "value": "Short-Horizon Intention Prediction",
+      "raw_hits": []
+    },
+    {
+      "name": "next_action: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current window at time t",
       "raw_hits": []
     },
     {
@@ -446,9 +452,9 @@
       "raw_hits": []
     },
     {
-      "name": "next_action: public_field_research_name_is_human_readable",
+      "name": "next_action: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Short-Horizon Intention Prediction",
+      "value": "current features -> future label shift -> classifier",
       "raw_hits": []
     },
     {
@@ -457,12 +463,6 @@
       "value": "Use the current window to guess the action that will happen shortly after it.",
       "raw_hits": []
     },
-    {
-      "name": "next_action: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "current features -> future label shift -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "next_action: known_task_family",
       "status": "pass",
@@ -540,15 +540,21 @@
       "observed": "hand_trajectory_forecast"
     },
     {
-      "name": "hand_trajectory_forecast: public_field_input_short_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current multimodal window",
+      "value": "Predict the future 3D left/right hand path from the current multimodal state.",
       "raw_hits": []
     },
     {
-      "name": "hand_trajectory_forecast: public_field_card_blurb_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict the future 3D left/right hand path from the current multimodal state.",
+      "value": "3D Hand Motion Forecasting",
+      "raw_hits": []
+    },
+    {
+      "name": "hand_trajectory_forecast: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current multimodal window",
       "raw_hits": []
     },
     {
@@ -564,9 +570,9 @@
       "raw_hits": []
     },
     {
-      "name": "hand_trajectory_forecast: public_field_research_name_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "3D Hand Motion Forecasting",
+      "value": "current features -> future mocap target -> regression head",
       "raw_hits": []
     },
     {
@@ -575,12 +581,6 @@
       "value": "Predict where the hands will move over the next few frames.",
       "raw_hits": []
     },
-    {
-      "name": "hand_trajectory_forecast: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "current features -> future mocap target -> regression head",
-      "raw_hits": []
-    },
     {
       "name": "hand_trajectory_forecast: known_task_family",
       "status": "pass",
@@ -658,15 +658,21 @@
       "observed": "contact_prediction"
     },
     {
-      "name": "contact_prediction: public_field_input_short_is_human_readable",
+      "name": "contact_prediction: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "non-contact, non-caption features",
+      "value": "Predict whether body or hand contact with the scene is occurring without leaking contact labels.",
       "raw_hits": []
     },
     {
-      "name": "contact_prediction: public_field_card_blurb_is_human_readable",
+      "name": "contact_prediction: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict whether body or hand contact with the scene is occurring without leaking contact labels.",
+      "value": "Human-Object Contact Prediction",
+      "raw_hits": []
+    },
+    {
+      "name": "contact_prediction: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "non-contact, non-caption features",
       "raw_hits": []
     },
     {
@@ -682,9 +688,9 @@
       "raw_hits": []
     },
     {
-      "name": "contact_prediction: public_field_research_name_is_human_readable",
+      "name": "contact_prediction: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Human-Object Contact Prediction",
+      "value": "feature filter -> contact target -> binary classifier",
       "raw_hits": []
     },
     {
@@ -693,12 +699,6 @@
       "value": "Predict whether the body or hand is in contact with something.",
       "raw_hits": []
     },
-    {
-      "name": "contact_prediction: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "feature filter -> contact target -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "contact_prediction: known_task_family",
       "status": "pass",
@@ -774,15 +774,21 @@
       "observed": "object_relevance"
     },
     {
-      "name": "object_relevance: public_field_input_short_is_human_readable",
+      "name": "object_relevance: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "non-caption multimodal features",
+      "value": "Infer which objects are relevant to the current manipulation window from non-caption features.",
       "raw_hits": []
     },
     {
-      "name": "object_relevance: public_field_card_blurb_is_human_readable",
+      "name": "object_relevance: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Infer which objects are relevant to the current manipulation window from non-caption features.",
+      "value": "Object-Centric Interaction Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "object_relevance: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "non-caption multimodal features",
       "raw_hits": []
     },
     {
@@ -798,9 +804,9 @@
       "raw_hits": []
     },
     {
-      "name": "object_relevance: public_field_research_name_is_human_readable",
+      "name": "object_relevance: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Object-Centric Interaction Recognition",
+      "value": "object vocabulary -> multi-hot labels -> sigmoid heads",
       "raw_hits": []
     },
     {
@@ -809,12 +815,6 @@
       "value": "Predict which objects matter in the current window.",
       "raw_hits": []
     },
-    {
-      "name": "object_relevance: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "object vocabulary -> multi-hot labels -> sigmoid heads",
-      "raw_hits": []
-    },
     {
       "name": "object_relevance: known_task_family",
       "status": "pass",
@@ -892,15 +892,21 @@
       "observed": "caption_grounding"
     },
     {
-      "name": "caption_grounding: public_field_input_short_is_human_readable",
+      "name": "caption_grounding: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "text-like query and candidate windows",
+      "value": "Retrieve the matching time window for an annotation-derived text query.",
       "raw_hits": []
     },
     {
-      "name": "caption_grounding: public_field_card_blurb_is_human_readable",
+      "name": "caption_grounding: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Retrieve the matching time window for an annotation-derived text query.",
+      "value": "Language-to-Moment Grounding",
+      "raw_hits": []
+    },
+    {
+      "name": "caption_grounding: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "text-like query and candidate windows",
       "raw_hits": []
     },
     {
@@ -916,9 +922,9 @@
       "raw_hits": []
     },
     {
-      "name": "caption_grounding: public_field_research_name_is_human_readable",
+      "name": "caption_grounding: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Language-to-Moment Grounding",
+      "value": "query features -> candidate index -> cosine ranker",
       "raw_hits": []
     },
     {
@@ -927,12 +933,6 @@
       "value": "Given a text-like query from annotation, find the matching time window.",
       "raw_hits": []
     },
-    {
-      "name": "caption_grounding: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "query features -> candidate index -> cosine ranker",
-      "raw_hits": []
-    },
     {
       "name": "caption_grounding: known_task_family",
       "status": "pass",
@@ -1008,15 +1008,21 @@
       "observed": "cross_modal_retrieval"
     },
     {
-      "name": "cross_modal_retrieval: public_field_input_short_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion/IMU/pose query; depth/video candidates",
+      "value": "Use motion, IMU, and camera-pose signals to retrieve the matching depth/video window.",
       "raw_hits": []
     },
     {
-      "name": "cross_modal_retrieval: public_field_card_blurb_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Use motion, IMU, and camera-pose signals to retrieve the matching depth/video window.",
+      "value": "Multimodal Representation Retrieval",
+      "raw_hits": []
+    },
+    {
+      "name": "cross_modal_retrieval: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion/IMU/pose query; depth/video candidates",
       "raw_hits": []
     },
     {
@@ -1032,9 +1038,9 @@
       "raw_hits": []
     },
     {
-      "name": "cross_modal_retrieval: public_field_research_name_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Multimodal Representation Retrieval",
+      "value": "modality split -> projection -> nearest-neighbor ranker",
       "raw_hits": []
     },
     {
@@ -1043,12 +1049,6 @@
       "value": "Use one group of modalities to retrieve the matching window from another group.",
       "raw_hits": []
     },
-    {
-      "name": "cross_modal_retrieval: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "modality split -> projection -> nearest-neighbor ranker",
-      "raw_hits": []
-    },
     {
       "name": "cross_modal_retrieval: known_task_family",
       "status": "pass",
@@ -1126,15 +1126,21 @@
       "observed": "modality_reconstruction"
     },
     {
-      "name": "modality_reconstruction: public_field_input_short_is_human_readable",
+      "name": "modality_reconstruction: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion, IMU, and camera/pose features",
+      "value": "Predict compressed depth/video feature vectors from motion, IMU, and camera-pose features.",
       "raw_hits": []
     },
     {
-      "name": "modality_reconstruction: public_field_card_blurb_is_human_readable",
+      "name": "modality_reconstruction: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict compressed depth/video feature vectors from motion, IMU, and camera-pose features.",
+      "value": "Modality Feature Reconstruction",
+      "raw_hits": []
+    },
+    {
+      "name": "modality_reconstruction: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion, IMU, and camera/pose features",
       "raw_hits": []
     },
     {
@@ -1150,9 +1156,9 @@
       "raw_hits": []
     },
     {
-      "name": "modality_reconstruction: public_field_research_name_is_human_readable",
+      "name": "modality_reconstruction: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Modality Feature Reconstruction",
+      "value": "source-target split -> scaler -> regression head",
       "raw_hits": []
     },
     {
@@ -1161,12 +1167,6 @@
       "value": "Predict one modality feature block from other modality blocks.",
       "raw_hits": []
     },
-    {
-      "name": "modality_reconstruction: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "source-target split -> scaler -> regression head",
-      "raw_hits": []
-    },
     {
       "name": "modality_reconstruction: known_task_family",
       "status": "pass",
@@ -1243,12 +1243,6 @@
       "status": "pass",
       "observed": "temporal_order"
     },
-    {
-      "name": "temporal_order: public_field_input_short_is_human_readable",
-      "status": "pass",
-      "value": "two adjacent windows plus difference vector",
-      "raw_hits": []
-    },
     {
       "name": "temporal_order: public_field_card_blurb_is_human_readable",
       "status": "pass",
@@ -1256,27 +1250,27 @@
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_display_name_is_human_readable",
+      "name": "temporal_order: public_field_research_name_is_human_readable",
       "status": "pass",
       "value": "Temporal Order Verification",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_output_short_is_human_readable",
+      "name": "temporal_order: public_field_input_short_is_human_readable",
       "status": "pass",
-      "value": "correct or reversed",
+      "value": "two adjacent windows plus difference vector",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_research_name_is_human_readable",
+      "name": "temporal_order: public_field_display_name_is_human_readable",
       "status": "pass",
       "value": "Temporal Order Verification",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_plain_goal_is_human_readable",
+      "name": "temporal_order: public_field_output_short_is_human_readable",
       "status": "pass",
-      "value": "Tell whether two nearby windows are in the correct time order.",
+      "value": "correct or reversed",
       "raw_hits": []
     },
     {
@@ -1285,6 +1279,12 @@
       "value": "pair builder -> feature combiner -> binary classifier",
       "raw_hits": []
     },
+    {
+      "name": "temporal_order: public_field_plain_goal_is_human_readable",
+      "status": "pass",
+      "value": "Tell whether two nearby windows are in the correct time order.",
+      "raw_hits": []
+    },
     {
       "name": "temporal_order: known_task_family",
       "status": "pass",
@@ -1360,15 +1360,21 @@
       "observed": "misalignment_detection"
     },
     {
-      "name": "misalignment_detection: public_field_input_short_is_human_readable",
+      "name": "misalignment_detection: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion-side and visual/depth-side feature groups",
+      "value": "Detect whether motion and visual/depth streams have been artificially shifted out of sync.",
       "raw_hits": []
     },
     {
-      "name": "misalignment_detection: public_field_card_blurb_is_human_readable",
+      "name": "misalignment_detection: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Detect whether motion and visual/depth streams have been artificially shifted out of sync.",
+      "value": "Cross-Modal Misalignment Detection",
+      "raw_hits": []
+    },
+    {
+      "name": "misalignment_detection: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion-side and visual/depth-side feature groups",
       "raw_hits": []
     },
     {
@@ -1384,9 +1390,9 @@
       "raw_hits": []
     },
     {
-      "name": "misalignment_detection: public_field_research_name_is_human_readable",
+      "name": "misalignment_detection: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Cross-Modal Misalignment Detection",
+      "value": "aligned/shifted pairs -> feature combiner -> binary classifier",
       "raw_hits": []
     },
     {
@@ -1395,12 +1401,6 @@
       "value": "Detect when modalities that should match are shifted out of sync.",
       "raw_hits": []
     },
-    {
-      "name": "misalignment_detection: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "aligned/shifted pairs -> feature combiner -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "misalignment_detection: known_task_family",
       "status": "pass",

docs/data/website_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:36:10+00:00",
+  "generated_at_utc": "2026-06-06T14:54:01+00:00",
   "docs_root": "docs",
   "site_base": "/ropedia-xperience-10m-task-suite/",
   "summary": {
@@ -251,7 +251,7 @@
     },
     {
       "path": "data/artifact_index.json",
-      "bytes": 37736,
+      "bytes": 39486,
       "top_level_type": "dict"
     },
     {
@@ -291,7 +291,7 @@
     },
     {
       "path": "data/mirror_parity.json",
-      "bytes": 111950,
+      "bytes": 126335,
       "top_level_type": "dict"
     },
     {
@@ -301,7 +301,7 @@
     },
     {
       "path": "data/omni_finetune_verified_result.json",
-      "bytes": 3145,
+      "bytes": 4142,
       "top_level_type": "dict"
     },
     {
@@ -321,7 +321,7 @@
     },
     {
       "path": "data/project_status.json",
-      "bytes": 10977,
+      "bytes": 11274,
       "top_level_type": "dict"
     },
     {

metrics/artifact_index.json

@@ -1,12 +1,12 @@
 {
   "title": "Ropedia Xperience-10M Task Suite Artifact Index",
-  "generated_at_utc": "2026-06-06T14:35:42+00:00",
+  "generated_at_utc": "2026-06-06T14:53:45+00:00",
   "status": "pass",
-  "artifact_count": 83,
+  "artifact_count": 86,
   "missing": [],
   "by_kind": {
     "project_path": 14,
-    "scaleup_contract": 6,
+    "scaleup_contract": 7,
     "project_scope": 1,
     "source_alignment": 5,
     "publication_workflow": 3,
@@ -28,7 +28,7 @@
     "onboarding_doc": 1,
     "generated_figure": 3,
     "generated_figure_assets": 1,
-    "scaleup_status": 2,
+    "scaleup_status": 4,
     "citation": 1,
     "license": 1
   },
@@ -63,8 +63,8 @@
       "surface": "repo_hf",
       "shows": "Gives a compact current-state table for first-pass readers.",
       "exists": true,
-      "bytes": 8534,
-      "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+      "bytes": 8805,
+      "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
     },
     {
       "id": "project_status_json",
@@ -74,8 +74,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable copy of the current project status for website and HF mirrors.",
       "exists": true,
-      "bytes": 10977,
-      "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+      "bytes": 11274,
+      "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
     },
     {
       "id": "research_roadmap",
@@ -187,6 +187,17 @@
       "bytes": 6519,
       "sha256": "a3773fc681e298325e2be80556d6be6e7e30b90ba22ee24b66633f07ff9c4ea4"
     },
+    {
+      "id": "qwen3_omni_error_analysis_script",
+      "title": "Qwen3-Omni held-out error-analysis script",
+      "path": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "kind": "scaleup_contract",
+      "surface": "repo_hf",
+      "shows": "Computes public-safe held-out error-analysis tables by episode, action family, train-seen status, required-modality state, and object category.",
+      "exists": true,
+      "bytes": 15676,
+      "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+    },
     {
       "id": "additional_development_directions",
       "title": "Additional development directions",
@@ -250,8 +261,8 @@
       "surface": "repo_hf",
       "shows": "Gives the human-readable map from project scope to data, tasks, platform mirrors, and scale-up status.",
       "exists": true,
-      "bytes": 15660,
-      "sha256": "a9ad335b82c35a5ac102428663ffae1c8798e90e45cc5e795c3a499b4563b417"
+      "bytes": 16318,
+      "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
     },
     {
       "id": "official_dataset_card_alignment",
@@ -695,8 +706,8 @@
       "surface": "repo_hf",
       "shows": "Generates the selective artifact catalog from local files.",
       "exists": true,
-      "bytes": 30785,
-      "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+      "bytes": 32191,
+      "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
     },
     {
       "id": "publication_audit",
@@ -731,7 +742,7 @@
       "volatile": true,
       "shows": "Confirms prepared GitHub/HF Space/artifact/model mirrors share the same critical data, figure, website HTML, and validator files.",
       "exists": true,
-      "bytes": 111950,
+      "bytes": 126335,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -933,6 +944,28 @@
       "bytes": 3076,
       "sha256": "23b87581cfc1d95b0af118a0dbb4e601f42fc6bad608759490e13a9a1ef73205"
     },
+    {
+      "id": "qwen3_omni_error_analysis_report",
+      "title": "Qwen3-Omni held-out error-analysis report",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "kind": "scaleup_status",
+      "surface": "repo_hf",
+      "shows": "Summarizes validation-aware Qwen3-Omni held-out failures by episode, action family, train-seen status, required-modality state, and object category.",
+      "exists": true,
+      "bytes": 3331,
+      "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+    },
+    {
+      "id": "qwen3_omni_error_analysis_json",
+      "title": "Qwen3-Omni held-out error-analysis JSON",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "kind": "scaleup_status",
+      "surface": "repo_hf",
+      "shows": "Machine-readable Qwen3-Omni held-out error analysis with grouped metrics and sanitized failure examples.",
+      "exists": true,
+      "bytes": 25202,
+      "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+    },
     {
       "id": "citation",
       "title": "Citation metadata",

metrics/mirror_parity.json

@@ -1,9 +1,9 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:37:36+00:00",
+  "generated_at_utc": "2026-06-06T14:56:44+00:00",
   "hf_root": "hf_publish",
   "summary": {
-    "group_count": 104,
+    "group_count": 114,
     "failure_count": 0,
     "failures_by_surface": {}
   },
@@ -102,27 +102,27 @@
       "local": {
         "path": "repo:docs/data/artifact_index.json",
         "exists": true,
-        "bytes": 37736,
-        "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+        "bytes": 39486,
+        "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
         },
         "hf_model": {
           "path": "hf_model:metrics/artifact_index.json",
           "exists": true,
-          "bytes": 37736,
-          "sha256": "f1d87cbabab02227b834ad333507af31a8ce309600f0e0427bb8cb59a26c3b71"
+          "bytes": 39486,
+          "sha256": "87782cd08bc1106d694a727e21333450d2965b48c48f500d1b6f4294d7b247d0"
         }
       },
       "failures": []
@@ -350,27 +350,27 @@
       "local": {
         "path": "repo:docs/data/omni_finetune_verified_result.json",
         "exists": true,
-        "bytes": 3145,
-        "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+        "bytes": 4142,
+        "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         },
         "hf_model": {
           "path": "hf_model:metrics/omni_finetune_verified_result.json",
           "exists": true,
-          "bytes": 3145,
-          "sha256": "37b001a24201ba56b327fa89f19792d64ebcdabc1faffa7e7bb4fd6b8323731a"
+          "bytes": 4142,
+          "sha256": "297aa6fc86bc09ba7968f3c5c2db265320c0613c5ec9a36701114ba451321b81"
         }
       },
       "failures": []
@@ -474,27 +474,27 @@
       "local": {
         "path": "repo:docs/data/project_status.json",
         "exists": true,
-        "bytes": 10977,
-        "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+        "bytes": 11274,
+        "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         },
         "hf_model": {
           "path": "hf_model:metrics/project_status.json",
           "exists": true,
-          "bytes": 10977,
-          "sha256": "2bb0639c137dfd6eddd337eb909292543ae2e72753dee398f8240ff35f6a3984"
+          "bytes": 11274,
+          "sha256": "ae2b2c520ab1e0553fa399439345edd87832fa5293d8c27ffe610ede5bfa1067"
         }
       },
       "failures": []
@@ -506,26 +506,26 @@
         "path": "repo:docs/data/publication_audit.json",
         "exists": true,
         "bytes": 7237,
-        "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+        "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
         },
         "hf_model": {
           "path": "hf_model:metrics/publication_audit.json",
           "exists": true,
           "bytes": 7237,
-          "sha256": "bfdfb04abf62dfb3ffa596f1d9ec58fc5bac633f6c1cfb1710d3988ef635cf03"
+          "sha256": "8a21c29d92f3a15b835c37d7784c17fada3edbda050515deed8e440535ed046d"
         }
       },
       "failures": []
@@ -816,26 +816,26 @@
         "path": "repo:docs/data/scope_claims_audit.json",
         "exists": true,
         "bytes": 20823,
-        "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+        "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
         },
         "hf_model": {
           "path": "hf_model:metrics/scope_claims_audit.json",
           "exists": true,
           "bytes": 20823,
-          "sha256": "7f01728415c9c54126eab25f2ce68e563b455f02d2bf10af514463c33bc0091e"
+          "sha256": "77402dc77c4ecf5cf1e68480ae2c9822a134ae7ef4a24a7b8b9008a2509c2fa3"
         }
       },
       "failures": []
@@ -940,26 +940,26 @@
         "path": "repo:docs/data/task_surface_integrity.json",
         "exists": true,
         "bytes": 45779,
-        "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+        "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
         },
         "hf_model": {
           "path": "hf_model:metrics/task_surface_integrity.json",
           "exists": true,
           "bytes": 45779,
-          "sha256": "1ae426aea9895c32912b2c9a0e519a55912222493d3c1d72e4785d71cd3b71cb"
+          "sha256": "8232e2bafa8b5157d97c018e41be5da3ec69ddb4d2020a0dcc7c6377c5575bb6"
         }
       },
       "failures": []
@@ -1002,26 +1002,26 @@
         "path": "repo:docs/data/website_integrity.json",
         "exists": true,
         "bytes": 15221,
-        "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+        "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:data/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:docs/data/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
         },
         "hf_model": {
           "path": "hf_model:metrics/website_integrity.json",
           "exists": true,
           "bytes": 15221,
-          "sha256": "08f9429aead121834f52fb108a35ff0933435d49064650b94b7ed84c1002182b"
+          "sha256": "dcbd09b4c4522770c43504c500eb653de706538516ee2ec72e491ffc3416c6e2"
         }
       },
       "failures": []
@@ -1723,6 +1723,31 @@
       },
       "failures": []
     },
+    {
+      "name": "scripts/omni/analyze_qwen3_omni_errors.py",
+      "status": "pass",
+      "local": {
+        "path": "repo:scripts/omni/analyze_qwen3_omni_errors.py",
+        "exists": true,
+        "bytes": 15676,
+        "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+      },
+      "mirrors": {
+        "hf_artifacts": {
+          "path": "hf_artifacts:scripts/omni/analyze_qwen3_omni_errors.py",
+          "exists": true,
+          "bytes": 15676,
+          "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+        },
+        "hf_model": {
+          "path": "hf_model:scripts/omni/analyze_qwen3_omni_errors.py",
+          "exists": true,
+          "bytes": 15676,
+          "sha256": "d4c7e46d9fbd5f9d84bc32374f457fd8c9d68c8faa39c77bc45770eb95d80337"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "scripts/audio_ablation_and_raw_upgrade.py",
       "status": "pass",
@@ -1754,21 +1779,21 @@
       "local": {
         "path": "repo:scripts/build_artifact_index.py",
         "exists": true,
-        "bytes": 30785,
-        "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+        "bytes": 32191,
+        "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 30785,
-          "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+          "bytes": 32191,
+          "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
         },
         "hf_model": {
           "path": "hf_model:scripts/build_artifact_index.py",
           "exists": true,
-          "bytes": 30785,
-          "sha256": "0c42b68e44e6a32b6b5161b47161adc5ccdb57567e1462e8271ea87af50ab92d"
+          "bytes": 32191,
+          "sha256": "4a105c732d2f6c54a78333d7f47e0139325ba638027e34e6acd929a90626b8e0"
         }
       },
       "failures": []
@@ -2054,21 +2079,21 @@
       "local": {
         "path": "repo:scripts/validate_mirror_parity.py",
         "exists": true,
-        "bytes": 12642,
-        "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+        "bytes": 13781,
+        "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
       },
       "mirrors": {
         "hf_artifacts": {
           "path": "hf_artifacts:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 12642,
-          "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+          "bytes": 13781,
+          "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
         },
         "hf_model": {
           "path": "hf_model:scripts/validate_mirror_parity.py",
           "exists": true,
-          "bytes": 12642,
-          "sha256": "17420a261d1327c0a8acb79adb75fc15217f117216eb74acf0cab3fa36de856c"
+          "bytes": 13781,
+          "sha256": "3659adf936b058617dde97ee4c424615a361e59f5ea74975116422dfe01768e8"
         }
       },
       "failures": []
@@ -2807,6 +2832,285 @@
       },
       "failures": []
     },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+        "exists": true,
+        "bytes": 3331,
+        "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+          "exists": true,
+          "bytes": 3331,
+          "sha256": "063fcc2ebd7b57ab5b281fd5e8edc629da4e1f4e5a708483ba27375d02af9467"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+        "exists": true,
+        "bytes": 25202,
+        "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+          "exists": true,
+          "bytes": 25202,
+          "sha256": "c2e4eaa686f5d9739a8d0bfd8ae51a453b94019489ed84a154e2bce2fa316ff5"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+        "exists": true,
+        "bytes": 2121,
+        "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2121,
+          "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2121,
+          "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+          "exists": true,
+          "bytes": 2121,
+          "sha256": "7f0bc74140f100b9fe444c38eb74d155605bfc5984f665e653a2cd34a5cb96bd"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+        "exists": true,
+        "bytes": 1320,
+        "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1320,
+          "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1320,
+          "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+          "exists": true,
+          "bytes": 1320,
+          "sha256": "e15bf22e96b887c4b00aeb8ba548f4fd72ea0aab0772cc59e9bdda517ad72430"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+        "exists": true,
+        "bytes": 572,
+        "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 572,
+          "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 572,
+          "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+          "exists": true,
+          "bytes": 572,
+          "sha256": "cb196616b6f073266087d8cb7182e36c0a761607f3082ad78c350fd99e1996e7"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+        "exists": true,
+        "bytes": 408,
+        "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 408,
+          "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 408,
+          "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+          "exists": true,
+          "bytes": 408,
+          "sha256": "6447cf285b466a914055adb0aef4f3d47bf82d33a277d8ca2e6f22c4f0f2a7f7"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+      "status": "pass",
+      "local": {
+        "path": "repo:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+        "exists": true,
+        "bytes": 1704,
+        "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1704,
+          "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1704,
+          "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+        },
+        "hf_model": {
+          "path": "hf_model:results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
+          "exists": true,
+          "bytes": 1704,
+          "sha256": "f9cbd5e566ef666fe2d1050cc5bdadc7967a2056bdaa1e2e9f88fb0c22ee0ef8"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "docs/ARTIFACT_GUIDE.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:ARTIFACT_GUIDE.md",
+        "exists": true,
+        "bytes": 16318,
+        "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        },
+        "hf_model": {
+          "path": "hf_model:ARTIFACT_GUIDE.md",
+          "exists": true,
+          "bytes": 16318,
+          "sha256": "cda5f4b5be4b7a2d26aff6ed7f930bfba13dfc463d533a9880193c0a0611b677"
+        }
+      },
+      "failures": []
+    },
+    {
+      "name": "docs/OMNI_MODEL_EXTENSION_CONTRACT.md",
+      "status": "pass",
+      "local": {
+        "path": "repo:OMNI_MODEL_EXTENSION_CONTRACT.md",
+        "exists": true,
+        "bytes": 8900,
+        "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+      },
+      "mirrors": {
+        "hf_space": {
+          "path": "hf_space:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        },
+        "hf_artifacts": {
+          "path": "hf_artifacts:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        },
+        "hf_model": {
+          "path": "hf_model:OMNI_MODEL_EXTENSION_CONTRACT.md",
+          "exists": true,
+          "bytes": 8900,
+          "sha256": "c4e51d0aa7536045c229418603a67c6b3c5f31c9d756ca7395cb0c9455f0ed6d"
+        }
+      },
+      "failures": []
+    },
     {
       "name": "docs/QUALITY_GATES.md",
       "status": "pass",
@@ -3061,27 +3365,27 @@
       "local": {
         "path": "repo:PROJECT_STATUS.md",
         "exists": true,
-        "bytes": 8534,
-        "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+        "bytes": 8805,
+        "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
       },
       "mirrors": {
         "hf_space": {
           "path": "hf_space:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         },
         "hf_artifacts": {
           "path": "hf_artifacts:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         },
         "hf_model": {
           "path": "hf_model:PROJECT_STATUS.md",
           "exists": true,
-          "bytes": 8534,
-          "sha256": "5eb48d489da7f005baab233a94c9d6b209eb1e9ffdb138c8e0e600ece9239a29"
+          "bytes": 8805,
+          "sha256": "4051b78674306078880de33a144a499144b2487b11455c70a364a94cefa035a7"
         }
       },
       "failures": []

metrics/omni_finetune_verified_result.json

@@ -67,7 +67,28 @@
     "audit_status": "pass",
     "contains_raw_xperience10m_data": false,
     "contains_qwen_base_weights": false,
-    "contains_lora_weights": false
+    "contains_lora_weights": false,
+    "error_analysis": {
+      "status": "pass",
+      "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+      "markdown_report": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+      "groupings": [
+        "episode",
+        "action_family",
+        "train_seen_status",
+        "required_modality_state",
+        "object_category"
+      ],
+      "key_readouts": {
+        "parsed_prediction_rate": 0.8772321428571429,
+        "weakest_action_family": "locomotion",
+        "weakest_action_family_samples": 23,
+        "weakest_action_family_parsed_prediction_rate": 0.2608695652173913,
+        "seen_action_exact_rate": 0.04580152671755725,
+        "unseen_action_exact_rate": 0.015772870662460567,
+        "required_modality_state": "rrd_missing_only_required_modalities_present"
+      }
+    }
   },
   "required_next_steps": [
     "Improve JSON-format reliability through prompt, decoding, constrained parsing, or target formatting changes.",

metrics/project_status.json

@@ -180,10 +180,12 @@
       "evidence": [
         "docs/data/omni_finetune_verified_result.json",
         "results/omni_finetune/verified_public/",
+        "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/",
         "scripts/omni/package_verified_omni_result.py",
-        "scripts/omni/audit_verified_omni_package.py"
+        "scripts/omni/audit_verified_omni_package.py",
+        "scripts/omni/analyze_qwen3_omni_errors.py"
       ],
-      "readout": "The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, and 448 test predictions. JSON validity is 87.50%, below the 98% target, so it is a stronger diagnostic baseline but not a strong model-quality result."
+      "readout": "The selected 96/16/16 episode split produced a validation-aware public-safe held-out package with 3,808 exported windows, 512 validation windows, 448 test predictions, and derived error-analysis tables by episode, action family, train-seen status, required-modality state, and object category. JSON validity is 87.50%, below the 98% target, so it is a diagnostic baseline but not a strong model-quality result."
     },
     {
       "area": "Raw Xperience-10M redistribution",

metrics/publication_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:38:05+00:00",
+  "generated_at_utc": "2026-06-06T14:54:02+00:00",
   "checks": [
     {
       "name": "required_publication_assets_present",
@@ -182,8 +182,8 @@
     "github_repo": {
       "root": "repo",
       "exists": true,
-      "file_count": 442,
-      "text_file_count": 372,
+      "file_count": 450,
+      "text_file_count": 380,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -193,8 +193,8 @@
     "hf_space_bundle": {
       "root": "hf_publish/space",
       "exists": true,
-      "file_count": 356,
-      "text_file_count": 286,
+      "file_count": 363,
+      "text_file_count": 293,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -204,8 +204,8 @@
     "hf_artifact_bundle": {
       "root": "hf_publish/artifacts",
       "exists": true,
-      "file_count": 514,
-      "text_file_count": 420,
+      "file_count": 522,
+      "text_file_count": 428,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -215,8 +215,8 @@
     "hf_model_bundle": {
       "root": "hf_publish/model",
       "exists": true,
-      "file_count": 701,
-      "text_file_count": 572,
+      "file_count": 709,
+      "text_file_count": 580,
       "largest_file": {
         "path": "pytorch_model.bin",
         "bytes": 93495480

metrics/scope_claims_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:35:59+00:00",
+  "generated_at_utc": "2026-06-06T14:54:01+00:00",
   "summary": {
     "qwen3_omni_verified_diagnostic_pilot": true,
     "dataset_manifest_num_episodes": 119,

metrics/task_surface_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:35:59+00:00",
+  "generated_at_utc": "2026-06-06T14:53:59+00:00",
   "summary": {
     "task_count": 12,
     "expected_task_count": 12,
@@ -64,15 +64,21 @@
       "observed": "timeline_action"
     },
     {
-      "name": "timeline_action: public_field_input_short_is_human_readable",
+      "name": "timeline_action: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "20-frame multimodal window",
+      "value": "Recognize the current manipulation action from synchronized visual, motion, inertial, pose, and annotation context.",
       "raw_hits": []
     },
     {
-      "name": "timeline_action: public_field_card_blurb_is_human_readable",
+      "name": "timeline_action: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Recognize the current manipulation action from synchronized visual, motion, inertial, pose, and annotation context.",
+      "value": "Egocentric Action Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "timeline_action: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "20-frame multimodal window",
       "raw_hits": []
     },
     {
@@ -88,9 +94,9 @@
       "raw_hits": []
     },
     {
-      "name": "timeline_action: public_field_research_name_is_human_readable",
+      "name": "timeline_action: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Egocentric Action Recognition",
+      "value": "window features -> action label builder -> classifier",
       "raw_hits": []
     },
     {
@@ -99,12 +105,6 @@
       "value": "Look at one short multimodal window and name what action is happening now.",
       "raw_hits": []
     },
-    {
-      "name": "timeline_action: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "window features -> action label builder -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "timeline_action: known_task_family",
       "status": "pass",
@@ -184,15 +184,21 @@
       "observed": "timeline_subtask"
     },
     {
-      "name": "timeline_subtask: public_field_input_short_is_human_readable",
+      "name": "timeline_subtask: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "20-frame multimodal window",
+      "value": "Recognize the broader activity stage so fine actions become a readable procedure timeline.",
       "raw_hits": []
     },
     {
-      "name": "timeline_subtask: public_field_card_blurb_is_human_readable",
+      "name": "timeline_subtask: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Recognize the broader activity stage so fine actions become a readable procedure timeline.",
+      "value": "Temporal Subtask Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "timeline_subtask: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "20-frame multimodal window",
       "raw_hits": []
     },
     {
@@ -208,9 +214,9 @@
       "raw_hits": []
     },
     {
-      "name": "timeline_subtask: public_field_research_name_is_human_readable",
+      "name": "timeline_subtask: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Temporal Subtask Recognition",
+      "value": "window features -> subtask label builder -> classifier",
       "raw_hits": []
     },
     {
@@ -219,12 +225,6 @@
       "value": "Predict the higher-level task stage for the current window.",
       "raw_hits": []
     },
-    {
-      "name": "timeline_subtask: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "window features -> subtask label builder -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "timeline_subtask: known_task_family",
       "status": "pass",
@@ -304,15 +304,21 @@
       "observed": "transition_detection"
     },
     {
-      "name": "transition_detection: public_field_input_short_is_human_readable",
+      "name": "transition_detection: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current window with boundary target",
+      "value": "Detect the local moment where the episode changes from one action segment to the next.",
       "raw_hits": []
     },
     {
-      "name": "transition_detection: public_field_card_blurb_is_human_readable",
+      "name": "transition_detection: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Detect the local moment where the episode changes from one action segment to the next.",
+      "value": "Temporal Action Segmentation",
+      "raw_hits": []
+    },
+    {
+      "name": "transition_detection: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current window with boundary target",
       "raw_hits": []
     },
     {
@@ -328,9 +334,9 @@
       "raw_hits": []
     },
     {
-      "name": "transition_detection: public_field_research_name_is_human_readable",
+      "name": "transition_detection: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Temporal Action Segmentation",
+      "value": "action changes -> boundary labels -> binary classifier",
       "raw_hits": []
     },
     {
@@ -339,12 +345,6 @@
       "value": "Detect whether the current window is near a boundary between actions.",
       "raw_hits": []
     },
-    {
-      "name": "transition_detection: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "action changes -> boundary labels -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "transition_detection: known_task_family",
       "status": "pass",
@@ -422,15 +422,21 @@
       "observed": "next_action"
     },
     {
-      "name": "next_action: public_field_input_short_is_human_readable",
+      "name": "next_action: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current window at time t",
+      "value": "Forecast the near-future action from the current observations only.",
       "raw_hits": []
     },
     {
-      "name": "next_action: public_field_card_blurb_is_human_readable",
+      "name": "next_action: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Forecast the near-future action from the current observations only.",
+      "value": "Short-Horizon Intention Prediction",
+      "raw_hits": []
+    },
+    {
+      "name": "next_action: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current window at time t",
       "raw_hits": []
     },
     {
@@ -446,9 +452,9 @@
       "raw_hits": []
     },
     {
-      "name": "next_action: public_field_research_name_is_human_readable",
+      "name": "next_action: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Short-Horizon Intention Prediction",
+      "value": "current features -> future label shift -> classifier",
       "raw_hits": []
     },
     {
@@ -457,12 +463,6 @@
       "value": "Use the current window to guess the action that will happen shortly after it.",
       "raw_hits": []
     },
-    {
-      "name": "next_action: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "current features -> future label shift -> classifier",
-      "raw_hits": []
-    },
     {
       "name": "next_action: known_task_family",
       "status": "pass",
@@ -540,15 +540,21 @@
       "observed": "hand_trajectory_forecast"
     },
     {
-      "name": "hand_trajectory_forecast: public_field_input_short_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "current multimodal window",
+      "value": "Predict the future 3D left/right hand path from the current multimodal state.",
       "raw_hits": []
     },
     {
-      "name": "hand_trajectory_forecast: public_field_card_blurb_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict the future 3D left/right hand path from the current multimodal state.",
+      "value": "3D Hand Motion Forecasting",
+      "raw_hits": []
+    },
+    {
+      "name": "hand_trajectory_forecast: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "current multimodal window",
       "raw_hits": []
     },
     {
@@ -564,9 +570,9 @@
       "raw_hits": []
     },
     {
-      "name": "hand_trajectory_forecast: public_field_research_name_is_human_readable",
+      "name": "hand_trajectory_forecast: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "3D Hand Motion Forecasting",
+      "value": "current features -> future mocap target -> regression head",
       "raw_hits": []
     },
     {
@@ -575,12 +581,6 @@
       "value": "Predict where the hands will move over the next few frames.",
       "raw_hits": []
     },
-    {
-      "name": "hand_trajectory_forecast: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "current features -> future mocap target -> regression head",
-      "raw_hits": []
-    },
     {
       "name": "hand_trajectory_forecast: known_task_family",
       "status": "pass",
@@ -658,15 +658,21 @@
       "observed": "contact_prediction"
     },
     {
-      "name": "contact_prediction: public_field_input_short_is_human_readable",
+      "name": "contact_prediction: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "non-contact, non-caption features",
+      "value": "Predict whether body or hand contact with the scene is occurring without leaking contact labels.",
       "raw_hits": []
     },
     {
-      "name": "contact_prediction: public_field_card_blurb_is_human_readable",
+      "name": "contact_prediction: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict whether body or hand contact with the scene is occurring without leaking contact labels.",
+      "value": "Human-Object Contact Prediction",
+      "raw_hits": []
+    },
+    {
+      "name": "contact_prediction: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "non-contact, non-caption features",
       "raw_hits": []
     },
     {
@@ -682,9 +688,9 @@
       "raw_hits": []
     },
     {
-      "name": "contact_prediction: public_field_research_name_is_human_readable",
+      "name": "contact_prediction: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Human-Object Contact Prediction",
+      "value": "feature filter -> contact target -> binary classifier",
       "raw_hits": []
     },
     {
@@ -693,12 +699,6 @@
       "value": "Predict whether the body or hand is in contact with something.",
       "raw_hits": []
     },
-    {
-      "name": "contact_prediction: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "feature filter -> contact target -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "contact_prediction: known_task_family",
       "status": "pass",
@@ -774,15 +774,21 @@
       "observed": "object_relevance"
     },
     {
-      "name": "object_relevance: public_field_input_short_is_human_readable",
+      "name": "object_relevance: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "non-caption multimodal features",
+      "value": "Infer which objects are relevant to the current manipulation window from non-caption features.",
       "raw_hits": []
     },
     {
-      "name": "object_relevance: public_field_card_blurb_is_human_readable",
+      "name": "object_relevance: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Infer which objects are relevant to the current manipulation window from non-caption features.",
+      "value": "Object-Centric Interaction Recognition",
+      "raw_hits": []
+    },
+    {
+      "name": "object_relevance: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "non-caption multimodal features",
       "raw_hits": []
     },
     {
@@ -798,9 +804,9 @@
       "raw_hits": []
     },
     {
-      "name": "object_relevance: public_field_research_name_is_human_readable",
+      "name": "object_relevance: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Object-Centric Interaction Recognition",
+      "value": "object vocabulary -> multi-hot labels -> sigmoid heads",
       "raw_hits": []
     },
     {
@@ -809,12 +815,6 @@
       "value": "Predict which objects matter in the current window.",
       "raw_hits": []
     },
-    {
-      "name": "object_relevance: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "object vocabulary -> multi-hot labels -> sigmoid heads",
-      "raw_hits": []
-    },
     {
       "name": "object_relevance: known_task_family",
       "status": "pass",
@@ -892,15 +892,21 @@
       "observed": "caption_grounding"
     },
     {
-      "name": "caption_grounding: public_field_input_short_is_human_readable",
+      "name": "caption_grounding: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "text-like query and candidate windows",
+      "value": "Retrieve the matching time window for an annotation-derived text query.",
       "raw_hits": []
     },
     {
-      "name": "caption_grounding: public_field_card_blurb_is_human_readable",
+      "name": "caption_grounding: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Retrieve the matching time window for an annotation-derived text query.",
+      "value": "Language-to-Moment Grounding",
+      "raw_hits": []
+    },
+    {
+      "name": "caption_grounding: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "text-like query and candidate windows",
       "raw_hits": []
     },
     {
@@ -916,9 +922,9 @@
       "raw_hits": []
     },
     {
-      "name": "caption_grounding: public_field_research_name_is_human_readable",
+      "name": "caption_grounding: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Language-to-Moment Grounding",
+      "value": "query features -> candidate index -> cosine ranker",
       "raw_hits": []
     },
     {
@@ -927,12 +933,6 @@
       "value": "Given a text-like query from annotation, find the matching time window.",
       "raw_hits": []
     },
-    {
-      "name": "caption_grounding: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "query features -> candidate index -> cosine ranker",
-      "raw_hits": []
-    },
     {
       "name": "caption_grounding: known_task_family",
       "status": "pass",
@@ -1008,15 +1008,21 @@
       "observed": "cross_modal_retrieval"
     },
     {
-      "name": "cross_modal_retrieval: public_field_input_short_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion/IMU/pose query; depth/video candidates",
+      "value": "Use motion, IMU, and camera-pose signals to retrieve the matching depth/video window.",
       "raw_hits": []
     },
     {
-      "name": "cross_modal_retrieval: public_field_card_blurb_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Use motion, IMU, and camera-pose signals to retrieve the matching depth/video window.",
+      "value": "Multimodal Representation Retrieval",
+      "raw_hits": []
+    },
+    {
+      "name": "cross_modal_retrieval: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion/IMU/pose query; depth/video candidates",
       "raw_hits": []
     },
     {
@@ -1032,9 +1038,9 @@
       "raw_hits": []
     },
     {
-      "name": "cross_modal_retrieval: public_field_research_name_is_human_readable",
+      "name": "cross_modal_retrieval: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Multimodal Representation Retrieval",
+      "value": "modality split -> projection -> nearest-neighbor ranker",
       "raw_hits": []
     },
     {
@@ -1043,12 +1049,6 @@
       "value": "Use one group of modalities to retrieve the matching window from another group.",
       "raw_hits": []
     },
-    {
-      "name": "cross_modal_retrieval: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "modality split -> projection -> nearest-neighbor ranker",
-      "raw_hits": []
-    },
     {
       "name": "cross_modal_retrieval: known_task_family",
       "status": "pass",
@@ -1126,15 +1126,21 @@
       "observed": "modality_reconstruction"
     },
     {
-      "name": "modality_reconstruction: public_field_input_short_is_human_readable",
+      "name": "modality_reconstruction: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion, IMU, and camera/pose features",
+      "value": "Predict compressed depth/video feature vectors from motion, IMU, and camera-pose features.",
       "raw_hits": []
     },
     {
-      "name": "modality_reconstruction: public_field_card_blurb_is_human_readable",
+      "name": "modality_reconstruction: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Predict compressed depth/video feature vectors from motion, IMU, and camera-pose features.",
+      "value": "Modality Feature Reconstruction",
+      "raw_hits": []
+    },
+    {
+      "name": "modality_reconstruction: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion, IMU, and camera/pose features",
       "raw_hits": []
     },
     {
@@ -1150,9 +1156,9 @@
       "raw_hits": []
     },
     {
-      "name": "modality_reconstruction: public_field_research_name_is_human_readable",
+      "name": "modality_reconstruction: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Modality Feature Reconstruction",
+      "value": "source-target split -> scaler -> regression head",
       "raw_hits": []
     },
     {
@@ -1161,12 +1167,6 @@
       "value": "Predict one modality feature block from other modality blocks.",
       "raw_hits": []
     },
-    {
-      "name": "modality_reconstruction: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "source-target split -> scaler -> regression head",
-      "raw_hits": []
-    },
     {
       "name": "modality_reconstruction: known_task_family",
       "status": "pass",
@@ -1243,12 +1243,6 @@
       "status": "pass",
       "observed": "temporal_order"
     },
-    {
-      "name": "temporal_order: public_field_input_short_is_human_readable",
-      "status": "pass",
-      "value": "two adjacent windows plus difference vector",
-      "raw_hits": []
-    },
     {
       "name": "temporal_order: public_field_card_blurb_is_human_readable",
       "status": "pass",
@@ -1256,27 +1250,27 @@
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_display_name_is_human_readable",
+      "name": "temporal_order: public_field_research_name_is_human_readable",
       "status": "pass",
       "value": "Temporal Order Verification",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_output_short_is_human_readable",
+      "name": "temporal_order: public_field_input_short_is_human_readable",
       "status": "pass",
-      "value": "correct or reversed",
+      "value": "two adjacent windows plus difference vector",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_research_name_is_human_readable",
+      "name": "temporal_order: public_field_display_name_is_human_readable",
       "status": "pass",
       "value": "Temporal Order Verification",
       "raw_hits": []
     },
     {
-      "name": "temporal_order: public_field_plain_goal_is_human_readable",
+      "name": "temporal_order: public_field_output_short_is_human_readable",
       "status": "pass",
-      "value": "Tell whether two nearby windows are in the correct time order.",
+      "value": "correct or reversed",
       "raw_hits": []
     },
     {
@@ -1285,6 +1279,12 @@
       "value": "pair builder -> feature combiner -> binary classifier",
       "raw_hits": []
     },
+    {
+      "name": "temporal_order: public_field_plain_goal_is_human_readable",
+      "status": "pass",
+      "value": "Tell whether two nearby windows are in the correct time order.",
+      "raw_hits": []
+    },
     {
       "name": "temporal_order: known_task_family",
       "status": "pass",
@@ -1360,15 +1360,21 @@
       "observed": "misalignment_detection"
     },
     {
-      "name": "misalignment_detection: public_field_input_short_is_human_readable",
+      "name": "misalignment_detection: public_field_card_blurb_is_human_readable",
       "status": "pass",
-      "value": "motion-side and visual/depth-side feature groups",
+      "value": "Detect whether motion and visual/depth streams have been artificially shifted out of sync.",
       "raw_hits": []
     },
     {
-      "name": "misalignment_detection: public_field_card_blurb_is_human_readable",
+      "name": "misalignment_detection: public_field_research_name_is_human_readable",
       "status": "pass",
-      "value": "Detect whether motion and visual/depth streams have been artificially shifted out of sync.",
+      "value": "Cross-Modal Misalignment Detection",
+      "raw_hits": []
+    },
+    {
+      "name": "misalignment_detection: public_field_input_short_is_human_readable",
+      "status": "pass",
+      "value": "motion-side and visual/depth-side feature groups",
       "raw_hits": []
     },
     {
@@ -1384,9 +1390,9 @@
       "raw_hits": []
     },
     {
-      "name": "misalignment_detection: public_field_research_name_is_human_readable",
+      "name": "misalignment_detection: public_field_process_short_is_human_readable",
       "status": "pass",
-      "value": "Cross-Modal Misalignment Detection",
+      "value": "aligned/shifted pairs -> feature combiner -> binary classifier",
       "raw_hits": []
     },
     {
@@ -1395,12 +1401,6 @@
       "value": "Detect when modalities that should match are shifted out of sync.",
       "raw_hits": []
     },
-    {
-      "name": "misalignment_detection: public_field_process_short_is_human_readable",
-      "status": "pass",
-      "value": "aligned/shifted pairs -> feature combiner -> binary classifier",
-      "raw_hits": []
-    },
     {
       "name": "misalignment_detection: known_task_family",
       "status": "pass",

metrics/website_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-06T14:36:10+00:00",
+  "generated_at_utc": "2026-06-06T14:54:01+00:00",
   "docs_root": "docs",
   "site_base": "/ropedia-xperience-10m-task-suite/",
   "summary": {
@@ -251,7 +251,7 @@
     },
     {
       "path": "data/artifact_index.json",
-      "bytes": 37736,
+      "bytes": 39486,
       "top_level_type": "dict"
     },
     {
@@ -291,7 +291,7 @@
     },
     {
       "path": "data/mirror_parity.json",
-      "bytes": 111950,
+      "bytes": 126335,
       "top_level_type": "dict"
     },
     {
@@ -301,7 +301,7 @@
     },
     {
       "path": "data/omni_finetune_verified_result.json",
-      "bytes": 3145,
+      "bytes": 4142,
       "top_level_type": "dict"
     },
     {
@@ -321,7 +321,7 @@
     },
     {
       "path": "data/project_status.json",
-      "bytes": 10977,
+      "bytes": 11274,
       "top_level_type": "dict"
     },
     {

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/PUBLIC_RESULT_SUMMARY.md

@@ -22,4 +22,22 @@
 
 Raw Xperience-10M files, base-model weights, adapter or checkpoint weights, full checkpoints, and large archives are not included.
 
+## Error Analysis
+
+The package includes a derived held-out error analysis under `analysis/`. It
+groups the 448 public prediction rows by episode, coarse action family,
+train-seen status, required-modality state, and object category.
+
+Key readouts:
+
+- Official JSON validity from `metrics.json`: `0.8750`
+- Parsed prediction rate from public rows: `0.8772`
+- Weakest action family by parsed prediction rate: `locomotion` with 23 rows and `0.2609`
+- Train-seen split: seen labels have `0.0458` action exact rate; unseen labels have `0.0158`
+- Required-modality state: all held-out rows have required modalities present, with only `visualization.rrd` absent
+
+Use `analysis/ERROR_ANALYSIS.md` and
+`analysis/error_analysis_summary.json` before planning the next
+structured-output pass.
+
 Use this package as the source for README, website, and Hugging Face updates.

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md

@@ -0,0 +1,78 @@
+# Qwen3-Omni Held-Out Error Analysis
+
+This report is computed from the verified public package predictions. It contains only derived metrics and sanitized examples.
+
+## Overall
+
+- Prediction rows: `448`
+- JSON validity from `metrics.json`: `0.8750`
+- Parsed prediction rate from public rows: `0.8772`
+- Action exact rate: `0.0246`
+- Subtask exact rate: `0.0067`
+- Contact exact rate: `0.6451`
+- Object F1: `0.2230`
+
+## Weakest Episode Groups
+
+| group | samples | parsed_prediction_rate | action_exact_rate | object_f1 |
+| --- | --- | --- | --- | --- |
+| 1796b943-caad-43c6-b9bd-80b8d601f37d__ep1 | 32 | 0.5625 | 0.0000 | 0.0459 |
+| 8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1 | 32 | 0.7500 | 0.0312 | 0.0942 |
+| 33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1 | 32 | 0.8438 | 0.0000 | 0.0529 |
+| b750fab3-7fbb-43a0-b451-c64c4d4a64da__ep1 | 32 | 0.8438 | 0.0000 | 0.2353 |
+| ba18b7c1-21ff-45da-8452-41acce7fc8de__ep2 | 32 | 0.8438 | 0.0000 | 0.2836 |
+| ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2 | 32 | 0.8438 | 0.0625 | 0.0746 |
+| b9dd769b-e31a-4fdb-945e-5a60db6487b0__ep2 | 32 | 0.8750 | 0.0312 | 0.3265 |
+| 4b02bb38-384a-438a-b5f9-6131d85c34b0__ep1 | 32 | 0.8750 | 0.0938 | 0.2830 |
+
+## Action Families
+
+| group | samples | parsed_prediction_rate | action_exact_rate | subtask_exact_rate | object_f1 |
+| --- | --- | --- | --- | --- | --- |
+| locomotion | 23 | 0.2609 | 0.0000 | 0.0000 | 0.0120 |
+| food_kitchen | 5 | 0.6000 | 0.2000 | 0.0000 | 0.2727 |
+| cleaning | 8 | 0.7500 | 0.0000 | 0.0000 | 0.0000 |
+| other | 94 | 0.8511 | 0.0000 | 0.0000 | 0.1910 |
+| phone_use | 51 | 0.9020 | 0.0588 | 0.0196 | 0.3501 |
+| paper_cardboard_craft | 142 | 0.9225 | 0.0282 | 0.0141 | 0.2308 |
+| small_object_sorting | 87 | 0.9655 | 0.0000 | 0.0000 | 0.2740 |
+| retail_stocking | 38 | 0.9737 | 0.0789 | 0.0000 | 0.1564 |
+
+## Train-Seen Split
+
+| group | samples | parsed_prediction_rate | action_exact_rate | next_action_exact_rate |
+| --- | --- | --- | --- | --- |
+| unseen_in_train | 317 | 0.8454 | 0.0158 | 0.0158 |
+| seen_in_train | 131 | 0.9542 | 0.0458 | 0.0458 |
+
+## Required-Modality State
+
+| group | samples | parsed_prediction_rate | action_exact_rate | object_f1 |
+| --- | --- | --- | --- | --- |
+| rrd_missing_only_required_modalities_present | 448 | 0.8772 | 0.0246 | 0.2230 |
+
+## Object Categories
+
+| group | samples | object_precision | object_recall | object_f1 |
+| --- | --- | --- | --- | --- |
+| furniture_room | 96 | 0.2534 | 0.2334 | 0.2430 |
+| other_object | 135 | 0.1372 | 0.1643 | 0.1495 |
+| food_kitchen | 56 | 0.2228 | 0.2000 | 0.2108 |
+| cleaning | 8 | 0.0400 | 0.0476 | 0.0435 |
+| phone_device | 162 | 0.3252 | 0.3132 | 0.3191 |
+| paper_cardboard | 261 | 0.2227 | 0.3234 | 0.2638 |
+| craft_small_object | 106 | 0.2266 | 0.2581 | 0.2413 |
+| retail_container | 101 | 0.2028 | 0.1752 | 0.1880 |
+
+## Interpretation
+
+The diagnostic pilot is dominated by invalid or weak structured outputs and exact-label failures. These tables identify where to tighten JSON constraints, action/subtask target formatting, object vocabularies, and missing-modality robustness before claiming stronger model quality.
+
+Generated files:
+
+- `error_analysis_summary.json`
+- `episode_error_analysis.csv`
+- `action_family_error_analysis.csv`
+- `train_seen_error_analysis.csv`
+- `missing_modality_error_analysis.csv`
+- `object_category_error_analysis.csv`

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv

@@ -0,0 +1,9 @@
+group,samples,parsed_prediction_rate,action_exact_rate,subtask_exact_rate,transition_exact_rate,next_action_exact_rate,contact_exact_rate,object_precision,object_recall,object_f1
+locomotion,23,0.2608695652173913,0.0,0.0,0.2608695652173913,0.0,0.08695652173913043,0.010752688172043012,0.0136986301369863,0.012048192771084338
+food_kitchen,5,0.6,0.2,0.0,0.6,0.2,0.2,0.375,0.21428571428571427,0.2727272727272727
+cleaning,8,0.75,0.0,0.0,0.625,0.0,0.625,0.0,0.0,0.0
+other,94,0.851063829787234,0.0,0.0,0.8085106382978723,0.0,0.6063829787234043,0.17220543806646527,0.21428571428571427,0.19095477386934673
+phone_use,51,0.9019607843137255,0.058823529411764705,0.0196078431372549,0.8431372549019608,0.058823529411764705,0.5686274509803921,0.35542168674698793,0.34502923976608185,0.3501483679525222
+paper_cardboard_craft,142,0.9225352112676056,0.028169014084507043,0.014084507042253521,0.9154929577464789,0.028169014084507043,0.8169014084507042,0.1853233830845771,0.3059548254620123,0.2308288148721921
+small_object_sorting,87,0.9655172413793104,0.0,0.0,0.9425287356321839,0.0,0.5747126436781609,0.26515151515151514,0.2834008097165992,0.27397260273972607
+retail_stocking,38,0.9736842105263158,0.07894736842105263,0.0,0.9473684210526315,0.07894736842105263,0.7631578947368421,0.15384615384615385,0.1590909090909091,0.1564245810055866

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv

@@ -0,0 +1,15 @@
+group,samples,parsed_prediction_rate,action_exact_rate,subtask_exact_rate,transition_exact_rate,next_action_exact_rate,contact_exact_rate,object_precision,object_recall,object_f1
+1796b943-caad-43c6-b9bd-80b8d601f37d__ep1,32,0.5625,0.0,0.0,0.5625,0.0,0.53125,0.045871559633027525,0.045871559633027525,0.045871559633027525
+8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1,32,0.75,0.03125,0.0,0.71875,0.03125,0.4375,0.08108108108108109,0.1125,0.09424083769633508
+33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1,32,0.84375,0.0,0.0,0.6875,0.0,0.53125,0.043859649122807015,0.06666666666666667,0.05291005291005291
+b750fab3-7fbb-43a0-b451-c64c4d4a64da__ep1,32,0.84375,0.0,0.0,0.84375,0.0,0.375,0.2153846153846154,0.25925925925925924,0.23529411764705882
+ba18b7c1-21ff-45da-8452-41acce7fc8de__ep2,32,0.84375,0.0,0.0,0.84375,0.0,0.75,0.3,0.2689655172413793,0.2836363636363637
+ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2,32,0.84375,0.0625,0.0625,0.84375,0.0625,0.75,0.04830917874396135,0.16393442622950818,0.07462686567164178
+b9dd769b-e31a-4fdb-945e-5a60db6487b0__ep2,32,0.875,0.03125,0.0,0.8125,0.03125,0.40625,0.30303030303030304,0.35398230088495575,0.32653061224489793
+4b02bb38-384a-438a-b5f9-6131d85c34b0__ep1,32,0.875,0.09375,0.03125,0.8125,0.09375,0.40625,0.2608695652173913,0.30927835051546393,0.2830188679245283
+9c553886-83c5-4dc4-be5c-dcb269b3a771__ep2,32,0.9375,0.0,0.0,0.9375,0.0,0.9375,0.21333333333333335,0.2831858407079646,0.24334600760456274
+5399ef86-4df9-49bc-809f-8f4f92f9e659__ep6,32,0.9375,0.0,0.0,0.90625,0.0,0.78125,0.027777777777777776,0.027777777777777776,0.027777777777777776
+b6579cb5-0a71-4ca6-8808-1e2700be05c7__ep3,32,0.96875,0.03125,0.0,0.9375,0.03125,0.96875,0.5130434782608696,0.4573643410852713,0.48360655737704916
+a1012a57-385e-45a9-8a59-694a26fe92a5__ep1,32,1.0,0.0,0.0,1.0,0.0,0.90625,0.1927710843373494,0.48484848484848486,0.27586206896551724
+877779cd-25f3-4293-a3c4-39067dd9558c__ep4,32,1.0,0.0,0.0,1.0,0.0,0.34375,0.3402061855670103,0.3548387096774194,0.3473684210526316
+34f07a04-eb37-45a3-95ec-189ed5f4a85b__ep5,32,1.0,0.09375,0.0,1.0,0.09375,0.90625,0.18840579710144928,0.18055555555555555,0.1843971631205674

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json

@@ -0,0 +1,667 @@
+{
+  "status": "pass",
+  "source_package": "xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval",
+  "source_prediction_rows": 448,
+  "metrics_json_validity_rate": 0.875,
+  "computed": {
+    "group": "overall",
+    "samples": 448,
+    "parsed_prediction_rate": 0.8772321428571429,
+    "action_exact_rate": 0.024553571428571428,
+    "subtask_exact_rate": 0.006696428571428571,
+    "transition_exact_rate": 0.8504464285714286,
+    "next_action_exact_rate": 0.024553571428571428,
+    "contact_exact_rate": 0.6450892857142857,
+    "object_precision": 0.19611111111111112,
+    "object_recall": 0.25841874084919475,
+    "object_f1": 0.22299431459254582
+  },
+  "worst_episode_groups": [
+    {
+      "group": "1796b943-caad-43c6-b9bd-80b8d601f37d__ep1",
+      "samples": 32,
+      "parsed_prediction_rate": 0.5625,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.5625,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.53125,
+      "object_precision": 0.045871559633027525,
+      "object_recall": 0.045871559633027525,
+      "object_f1": 0.045871559633027525
+    },
+    {
+      "group": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "samples": 32,
+      "parsed_prediction_rate": 0.75,
+      "action_exact_rate": 0.03125,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.71875,
+      "next_action_exact_rate": 0.03125,
+      "contact_exact_rate": 0.4375,
+      "object_precision": 0.08108108108108109,
+      "object_recall": 0.1125,
+      "object_f1": 0.09424083769633508
+    },
+    {
+      "group": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1",
+      "samples": 32,
+      "parsed_prediction_rate": 0.84375,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.6875,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.53125,
+      "object_precision": 0.043859649122807015,
+      "object_recall": 0.06666666666666667,
+      "object_f1": 0.05291005291005291
+    },
+    {
+      "group": "b750fab3-7fbb-43a0-b451-c64c4d4a64da__ep1",
+      "samples": 32,
+      "parsed_prediction_rate": 0.84375,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.84375,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.375,
+      "object_precision": 0.2153846153846154,
+      "object_recall": 0.25925925925925924,
+      "object_f1": 0.23529411764705882
+    },
+    {
+      "group": "ba18b7c1-21ff-45da-8452-41acce7fc8de__ep2",
+      "samples": 32,
+      "parsed_prediction_rate": 0.84375,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.84375,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.75,
+      "object_precision": 0.3,
+      "object_recall": 0.2689655172413793,
+      "object_f1": 0.2836363636363637
+    },
+    {
+      "group": "ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2",
+      "samples": 32,
+      "parsed_prediction_rate": 0.84375,
+      "action_exact_rate": 0.0625,
+      "subtask_exact_rate": 0.0625,
+      "transition_exact_rate": 0.84375,
+      "next_action_exact_rate": 0.0625,
+      "contact_exact_rate": 0.75,
+      "object_precision": 0.04830917874396135,
+      "object_recall": 0.16393442622950818,
+      "object_f1": 0.07462686567164178
+    },
+    {
+      "group": "b9dd769b-e31a-4fdb-945e-5a60db6487b0__ep2",
+      "samples": 32,
+      "parsed_prediction_rate": 0.875,
+      "action_exact_rate": 0.03125,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.8125,
+      "next_action_exact_rate": 0.03125,
+      "contact_exact_rate": 0.40625,
+      "object_precision": 0.30303030303030304,
+      "object_recall": 0.35398230088495575,
+      "object_f1": 0.32653061224489793
+    },
+    {
+      "group": "4b02bb38-384a-438a-b5f9-6131d85c34b0__ep1",
+      "samples": 32,
+      "parsed_prediction_rate": 0.875,
+      "action_exact_rate": 0.09375,
+      "subtask_exact_rate": 0.03125,
+      "transition_exact_rate": 0.8125,
+      "next_action_exact_rate": 0.09375,
+      "contact_exact_rate": 0.40625,
+      "object_precision": 0.2608695652173913,
+      "object_recall": 0.30927835051546393,
+      "object_f1": 0.2830188679245283
+    }
+  ],
+  "action_family_groups": [
+    {
+      "group": "locomotion",
+      "samples": 23,
+      "parsed_prediction_rate": 0.2608695652173913,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.2608695652173913,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.08695652173913043,
+      "object_precision": 0.010752688172043012,
+      "object_recall": 0.0136986301369863,
+      "object_f1": 0.012048192771084338
+    },
+    {
+      "group": "food_kitchen",
+      "samples": 5,
+      "parsed_prediction_rate": 0.6,
+      "action_exact_rate": 0.2,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.6,
+      "next_action_exact_rate": 0.2,
+      "contact_exact_rate": 0.2,
+      "object_precision": 0.375,
+      "object_recall": 0.21428571428571427,
+      "object_f1": 0.2727272727272727
+    },
+    {
+      "group": "cleaning",
+      "samples": 8,
+      "parsed_prediction_rate": 0.75,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.625,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.625,
+      "object_precision": 0.0,
+      "object_recall": 0.0,
+      "object_f1": 0.0
+    },
+    {
+      "group": "other",
+      "samples": 94,
+      "parsed_prediction_rate": 0.851063829787234,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.8085106382978723,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.6063829787234043,
+      "object_precision": 0.17220543806646527,
+      "object_recall": 0.21428571428571427,
+      "object_f1": 0.19095477386934673
+    },
+    {
+      "group": "phone_use",
+      "samples": 51,
+      "parsed_prediction_rate": 0.9019607843137255,
+      "action_exact_rate": 0.058823529411764705,
+      "subtask_exact_rate": 0.0196078431372549,
+      "transition_exact_rate": 0.8431372549019608,
+      "next_action_exact_rate": 0.058823529411764705,
+      "contact_exact_rate": 0.5686274509803921,
+      "object_precision": 0.35542168674698793,
+      "object_recall": 0.34502923976608185,
+      "object_f1": 0.3501483679525222
+    },
+    {
+      "group": "paper_cardboard_craft",
+      "samples": 142,
+      "parsed_prediction_rate": 0.9225352112676056,
+      "action_exact_rate": 0.028169014084507043,
+      "subtask_exact_rate": 0.014084507042253521,
+      "transition_exact_rate": 0.9154929577464789,
+      "next_action_exact_rate": 0.028169014084507043,
+      "contact_exact_rate": 0.8169014084507042,
+      "object_precision": 0.1853233830845771,
+      "object_recall": 0.3059548254620123,
+      "object_f1": 0.2308288148721921
+    },
+    {
+      "group": "small_object_sorting",
+      "samples": 87,
+      "parsed_prediction_rate": 0.9655172413793104,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.9425287356321839,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.5747126436781609,
+      "object_precision": 0.26515151515151514,
+      "object_recall": 0.2834008097165992,
+      "object_f1": 0.27397260273972607
+    },
+    {
+      "group": "retail_stocking",
+      "samples": 38,
+      "parsed_prediction_rate": 0.9736842105263158,
+      "action_exact_rate": 0.07894736842105263,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.9473684210526315,
+      "next_action_exact_rate": 0.07894736842105263,
+      "contact_exact_rate": 0.7631578947368421,
+      "object_precision": 0.15384615384615385,
+      "object_recall": 0.1590909090909091,
+      "object_f1": 0.1564245810055866
+    }
+  ],
+  "train_seen_groups": [
+    {
+      "group": "unseen_in_train",
+      "samples": 317,
+      "parsed_prediction_rate": 0.8454258675078864,
+      "action_exact_rate": 0.015772870662460567,
+      "subtask_exact_rate": 0.006309148264984227,
+      "transition_exact_rate": 0.8233438485804416,
+      "next_action_exact_rate": 0.015772870662460567,
+      "contact_exact_rate": 0.6151419558359621,
+      "object_precision": 0.15804806991988346,
+      "object_recall": 0.23183760683760685,
+      "object_f1": 0.18796015591165008
+    },
+    {
+      "group": "seen_in_train",
+      "samples": 131,
+      "parsed_prediction_rate": 0.9541984732824428,
+      "action_exact_rate": 0.04580152671755725,
+      "subtask_exact_rate": 0.007633587786259542,
+      "transition_exact_rate": 0.916030534351145,
+      "next_action_exact_rate": 0.04580152671755725,
+      "contact_exact_rate": 0.7175572519083969,
+      "object_precision": 0.3185011709601874,
+      "object_recall": 0.31627906976744186,
+      "object_f1": 0.3173862310385064
+    }
+  ],
+  "missing_modality_groups": [
+    {
+      "group": "rrd_missing_only_required_modalities_present",
+      "samples": 448,
+      "parsed_prediction_rate": 0.8772321428571429,
+      "action_exact_rate": 0.024553571428571428,
+      "subtask_exact_rate": 0.006696428571428571,
+      "transition_exact_rate": 0.8504464285714286,
+      "next_action_exact_rate": 0.024553571428571428,
+      "contact_exact_rate": 0.6450892857142857,
+      "object_precision": 0.19611111111111112,
+      "object_recall": 0.25841874084919475,
+      "object_f1": 0.22299431459254582
+    }
+  ],
+  "object_category_groups": [
+    {
+      "group": "furniture_room",
+      "samples": 96,
+      "parsed_prediction_rate": 0.71875,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.7083333333333334,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.4166666666666667,
+      "object_precision": 0.2534246575342466,
+      "object_recall": 0.2334384858044164,
+      "object_f1": 0.24302134646962234
+    },
+    {
+      "group": "other_object",
+      "samples": 135,
+      "parsed_prediction_rate": 0.7925925925925926,
+      "action_exact_rate": 0.02962962962962963,
+      "subtask_exact_rate": 0.007407407407407408,
+      "transition_exact_rate": 0.762962962962963,
+      "next_action_exact_rate": 0.02962962962962963,
+      "contact_exact_rate": 0.6,
+      "object_precision": 0.13717693836978131,
+      "object_recall": 0.16428571428571428,
+      "object_f1": 0.1495124593716143
+    },
+    {
+      "group": "food_kitchen",
+      "samples": 56,
+      "parsed_prediction_rate": 0.8571428571428571,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.8214285714285714,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.7678571428571429,
+      "object_precision": 0.22277227722772278,
+      "object_recall": 0.2,
+      "object_f1": 0.2107728337236534
+    },
+    {
+      "group": "cleaning",
+      "samples": 8,
+      "parsed_prediction_rate": 0.875,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.875,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 0.625,
+      "object_precision": 0.04,
+      "object_recall": 0.047619047619047616,
+      "object_f1": 0.043478260869565216
+    },
+    {
+      "group": "phone_device",
+      "samples": 162,
+      "parsed_prediction_rate": 0.9074074074074074,
+      "action_exact_rate": 0.024691358024691357,
+      "subtask_exact_rate": 0.006172839506172839,
+      "transition_exact_rate": 0.8703703703703703,
+      "next_action_exact_rate": 0.024691358024691357,
+      "contact_exact_rate": 0.5864197530864198,
+      "object_precision": 0.32521739130434785,
+      "object_recall": 0.3132328308207705,
+      "object_f1": 0.31911262798634815
+    },
+    {
+      "group": "paper_cardboard",
+      "samples": 261,
+      "parsed_prediction_rate": 0.9080459770114943,
+      "action_exact_rate": 0.034482758620689655,
+      "subtask_exact_rate": 0.011494252873563218,
+      "transition_exact_rate": 0.8888888888888888,
+      "next_action_exact_rate": 0.034482758620689655,
+      "contact_exact_rate": 0.7203065134099617,
+      "object_precision": 0.22274881516587677,
+      "object_recall": 0.32339449541284404,
+      "object_f1": 0.2637979420018709
+    },
+    {
+      "group": "craft_small_object",
+      "samples": 106,
+      "parsed_prediction_rate": 0.9339622641509434,
+      "action_exact_rate": 0.02830188679245283,
+      "subtask_exact_rate": 0.009433962264150943,
+      "transition_exact_rate": 0.9150943396226415,
+      "next_action_exact_rate": 0.02830188679245283,
+      "contact_exact_rate": 0.5,
+      "object_precision": 0.22662889518413598,
+      "object_recall": 0.25806451612903225,
+      "object_f1": 0.24132730015082954
+    },
+    {
+      "group": "retail_container",
+      "samples": 101,
+      "parsed_prediction_rate": 0.9405940594059405,
+      "action_exact_rate": 0.0297029702970297,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.9108910891089109,
+      "next_action_exact_rate": 0.0297029702970297,
+      "contact_exact_rate": 0.7722772277227723,
+      "object_precision": 0.20279720279720279,
+      "object_recall": 0.17522658610271905,
+      "object_f1": 0.18800648298217182
+    },
+    {
+      "group": "tool_stationery",
+      "samples": 138,
+      "parsed_prediction_rate": 0.9565217391304348,
+      "action_exact_rate": 0.014492753623188406,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 0.9347826086956522,
+      "next_action_exact_rate": 0.014492753623188406,
+      "contact_exact_rate": 0.8043478260869565,
+      "object_precision": 0.27906976744186046,
+      "object_recall": 0.3894523326572008,
+      "object_f1": 0.32514817950889074
+    },
+    {
+      "group": "no_object_label",
+      "samples": 2,
+      "parsed_prediction_rate": 1.0,
+      "action_exact_rate": 0.0,
+      "subtask_exact_rate": 0.0,
+      "transition_exact_rate": 1.0,
+      "next_action_exact_rate": 0.0,
+      "contact_exact_rate": 1.0,
+      "object_precision": 0.0,
+      "object_recall": 0.0,
+      "object_f1": 0.0
+    }
+  ],
+  "invalid_json_examples": [
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:0",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Hold smartphone",
+      "raw_prediction_prefix": "{\"action\": \"Pour liquid into bowl\", \"contact\": \"yes\", \"evidence_window\": {\"end_frame\": 19, \"start_frame\": 0}, \"next_action\": \"Pour liquid into bowl\", \"objects\": [\"bottle\", \"bowl\", \"kitchen counter\", \"kitchen sink\", \"kitchen tap\", \"kitchen c"
+    },
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:90",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Move towards the stove",
+      "raw_prediction_prefix": "{\"action\": \"Walk through workspace\", \"contact\": \"no\", \"evidence_window\": {\"end_frame\": 1819, \"start_frame\": 1800}, \"next_action\": \"Walk through workspace\", \"objects\": [\"kitchen counter\", \"kitchen sink\", \"kitchen cabinets\", \"kitchen applianc"
+    },
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:100",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Open stove pot lid",
+      "raw_prediction_prefix": "{\"action\": \"Approach packing area\", \"contact\": \"no\", \"evidence_window\": {\"end_frame\": 2019, \"start_frame\": 2000}, \"next_action\": \"Approach packing area\", \"objects\": [\"kitchen counter\", \"kitchen sink\", \"kitchen cabinet\", \"kitchen utensils\", "
+    },
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:126",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Closing the door",
+      "raw_prediction_prefix": "{\"action\": \"Close door\", \"contact\": \"yes\", \"evidence_window\": {\"end_frame\": 2539, \"start_frame\": 2520}, \"next_action\": \"Close door\", \"objects\": [\"door\", \"door handle\", \"plastic bag\", \"blue bucket\", \"red bucket\", \"blue container\", \"red conta"
+    },
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:127",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Closing the door",
+      "raw_prediction_prefix": "{\"action\": \"Close the door\", \"contact\": \"yes\", \"evidence_window\": {\"end_frame\": 2559, \"start_frame\": 2540}, \"next_action\": \"Close the door\", \"objects\": [\"door\", \"door handle\", \"door knob\", \"door frame\", \"door hinge\", \"door latch\", \"door loc"
+    },
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:134",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Picking up bottle",
+      "raw_prediction_prefix": "{\"action\": \"Carry cereal boxes\", \"contact\": \"yes\", \"evidence_window\": {\"end_frame\": 2699, \"start_frame\": 2680}, \"next_action\": \"Carry cereal boxes\", \"objects\": [\"cereal boxes\", \"door\", \"washing machine\", \"laundry basket\", \"laundry detergent"
+    },
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:140",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Wipe kitchen counter",
+      "raw_prediction_prefix": "{\"action\": \"Walk through workspace\", \"contact\": \"no\", \"evidence_window\": {\"end_frame\": 2819, \"start_frame\": 2800}, \"next_action\": \"Walk through workspace\", \"objects\": [\"laundry machine\", \"laundry basket\", \"laundry detergent\", \"laundry soap\""
+    },
+    {
+      "id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1:qa:170",
+      "episode_id": "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1",
+      "true_action": "Reach for cleaning supplies",
+      "raw_prediction_prefix": "{\"action\": \"Close cabinet door\", \"contact\": \"yes\", \"evidence_window\": {\"end_frame\": 3419, \"start_frame\": 3400}, \"next_action\": \"Close cabinet door\", \"objects\": [\"cabinet\", \"cabinet door\", \"cabinet handle\", \"cabinet shelf\", \"cabinet drawer\","
+    },
+    {
+      "id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1:qa:0",
+      "episode_id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1",
+      "true_action": "Observe and walk through store",
+      "raw_prediction_prefix": "{\"action\": \"Walk through workspace\", \"contact\": \"no\", \"evidence_window\": {\"end_frame\": 19, \"start_frame\": 0}, \"next_action\": \"Walk through workspace\", \"objects\": [\"person\", \"mirror\", \"desk\", \"chair\", \"box\", \"shelf\", \"door\", \"wall\", \"floor\","
+    },
+    {
+      "id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1:qa:1",
+      "episode_id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1",
+      "true_action": "Observe and walk through store",
+      "raw_prediction_prefix": "{\"action\": \"Walk through workspace\", \"contact\": \"no\", \"evidence_window\": {\"end_frame\": 39, \"start_frame\": 20}, \"next_action\": \"Walk through workspace\", \"objects\": [\"person\", \"desk\", \"chair\", \"wall\", \"ceiling\", \"floor\", \"box\", \"cardboard\", \""
+    },
+    {
+      "id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1:qa:50",
+      "episode_id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1",
+      "true_action": "Walk towards shelves",
+      "raw_prediction_prefix": "{\"action\": \"Walk through workspace\", \"contact\": \"no\", \"evidence_window\": {\"end_frame\": 1019, \"start_frame\": 1000}, \"next_action\": \"Walk through workspace\", \"objects\": [\"person\", \"cardboard\", \"shelf\", \"door\", \"box\", \"jar\", \"lantern\", \"light\""
+    },
+    {
+      "id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1:qa:59",
+      "episode_id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1",
+      "true_action": "Observe workspace",
+      "raw_prediction_prefix": "{\"action\": \"Walk through workspace\", \"contact\": \"no\", \"evidence_window\": {\"end_frame\": 1199, \"start_frame\": 1180}, \"next_action\": \"Walk through workspace\", \"objects\": [\"cardboard\", \"cardboard box\", \"cardboard pieces\", \"cardboard sheet\", \"ca"
+    }
+  ],
+  "object_overgeneration_examples": [
+    {
+      "id": "a1012a57-385e-45a9-8a59-694a26fe92a5__ep1:qa:19",
+      "episode_id": "a1012a57-385e-45a9-8a59-694a26fe92a5__ep1",
+      "true_action": "Start cutting",
+      "predicted_object_count": 175,
+      "first_predicted_objects": [
+        "cardboard",
+        "cardboard box",
+        "cardboard pieces",
+        "cardboard sheet",
+        "cardboard square",
+        "cardboard tray",
+        "cardboard tube",
+        "utility knife",
+        "scissors",
+        "ruler",
+        "pen",
+        "marker",
+        "box",
+        "container",
+        "plastic container",
+        "tin can",
+        "jar",
+        "canned food",
+        "canned goods",
+        "canned product"
+      ]
+    },
+    {
+      "id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1:qa:70",
+      "episode_id": "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1",
+      "true_action": "Reach for wire hangers",
+      "predicted_object_count": 53,
+      "first_predicted_objects": [
+        "cardboard",
+        "cardboard box",
+        "cardboard pieces",
+        "cardboard shapes",
+        "cardboard squares",
+        "cardboard tray",
+        "cardboard tube",
+        "cardboard pieces",
+        "cardboard shapes",
+        "cardboard squares",
+        "cardboard tray",
+        "cardboard tube",
+        "blue foam pieces",
+        "blue foam sheet",
+        "blue product box",
+        "blue strip",
+        "canned food",
+        "canned goods",
+        "canned items",
+        "cans"
+      ]
+    },
+    {
+      "id": "ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2:qa:30",
+      "episode_id": "ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2",
+      "true_action": "Grasp lantern",
+      "predicted_object_count": 119,
+      "first_predicted_objects": [
+        "jar",
+        "red bowl",
+        "cardboard box",
+        "white paper",
+        "black bag",
+        "white bag",
+        "plastic bag",
+        "cardboard pieces",
+        "cardboard tray",
+        "cardboard sheet",
+        "cardboard shape",
+        "cardboard tube",
+        "cardboard strip",
+        "cardboard pattern",
+        "cardboard cutout",
+        "cardboard square",
+        "cardboard stack",
+        "plastic container",
+        "canned food",
+        "tin can"
+      ]
+    },
+    {
+      "id": "ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2:qa:176",
+      "episode_id": "ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2",
+      "true_action": "Release lantern",
+      "predicted_object_count": 205,
+      "first_predicted_objects": [
+        "jar",
+        "gift box",
+        "cardboard",
+        "paper lantern",
+        "plastic bag",
+        "plastic container",
+        "shopping bag",
+        "cardboard box",
+        "cardboard piece",
+        "cardboard tray",
+        "cardboard sheet",
+        "cardboard shape",
+        "cardboard pattern",
+        "cardboard square",
+        "cardboard strip",
+        "cardboard tube",
+        "cardboard piece",
+        "cardboard cutout",
+        "cardboard pattern piece",
+        "box"
+      ]
+    },
+    {
+      "id": "1796b943-caad-43c6-b9bd-80b8d601f37d__ep1:qa:40",
+      "episode_id": "1796b943-caad-43c6-b9bd-80b8d601f37d__ep1",
+      "true_action": "Move through the training room",
+      "predicted_object_count": 108,
+      "first_predicted_objects": [
+        "people",
+        "office chairs",
+        "desk",
+        "computer",
+        "laptop",
+        "office supplies",
+        "whiteboard",
+        "door",
+        "window",
+        "light fixture",
+        "wall",
+        "floor",
+        "box",
+        "cardboard",
+        "paper",
+        "plastic container",
+        "jar",
+        "bottle",
+        "canned food",
+        "snack package"
+      ]
+    }
+  ],
+  "modality_missing_by_episode": {
+    "8a8e1b3c-607e-4ada-b3fd-fa639727e92c__ep1": [
+      "visualization.rrd"
+    ],
+    "a1012a57-385e-45a9-8a59-694a26fe92a5__ep1": [
+      "visualization.rrd"
+    ],
+    "33f7ae08-ac1d-4321-9cb9-eca79016b359__ep1": [
+      "visualization.rrd"
+    ],
+    "9c553886-83c5-4dc4-be5c-dcb269b3a771__ep2": [
+      "visualization.rrd"
+    ],
+    "34f07a04-eb37-45a3-95ec-189ed5f4a85b__ep5": [
+      "visualization.rrd"
+    ],
+    "b9dd769b-e31a-4fdb-945e-5a60db6487b0__ep2": [
+      "visualization.rrd"
+    ],
+    "ba045ed4-ef25-404d-b756-8dcbd45b18fa__ep2": [
+      "visualization.rrd"
+    ],
+    "4b02bb38-384a-438a-b5f9-6131d85c34b0__ep1": [
+      "visualization.rrd"
+    ],
+    "5399ef86-4df9-49bc-809f-8f4f92f9e659__ep6": [
+      "visualization.rrd"
+    ],
+    "b750fab3-7fbb-43a0-b451-c64c4d4a64da__ep1": [
+      "visualization.rrd"
+    ],
+    "877779cd-25f3-4293-a3c4-39067dd9558c__ep4": [
+      "visualization.rrd"
+    ],
+    "1796b943-caad-43c6-b9bd-80b8d601f37d__ep1": [
+      "visualization.rrd"
+    ],
+    "ba18b7c1-21ff-45da-8452-41acce7fc8de__ep2": [
+      "visualization.rrd"
+    ],
+    "b6579cb5-0a71-4ca6-8808-1e2700be05c7__ep3": [
+      "visualization.rrd"
+    ]
+  },
+  "interpretation": "The diagnostic pilot is dominated by invalid or weak structured outputs and exact-label failures. These tables identify where to tighten JSON constraints, action/subtask target formatting, object vocabularies, and missing-modality robustness before claiming stronger model quality."
+}

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv

@@ -0,0 +1,2 @@
+group,samples,parsed_prediction_rate,action_exact_rate,subtask_exact_rate,transition_exact_rate,next_action_exact_rate,contact_exact_rate,object_precision,object_recall,object_f1
+rrd_missing_only_required_modalities_present,448,0.8772321428571429,0.024553571428571428,0.006696428571428571,0.8504464285714286,0.024553571428571428,0.6450892857142857,0.19611111111111112,0.25841874084919475,0.22299431459254582

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv

@@ -0,0 +1,11 @@
+group,samples,parsed_prediction_rate,action_exact_rate,subtask_exact_rate,transition_exact_rate,next_action_exact_rate,contact_exact_rate,object_precision,object_recall,object_f1
+furniture_room,96,0.71875,0.0,0.0,0.7083333333333334,0.0,0.4166666666666667,0.2534246575342466,0.2334384858044164,0.24302134646962234
+other_object,135,0.7925925925925926,0.02962962962962963,0.007407407407407408,0.762962962962963,0.02962962962962963,0.6,0.13717693836978131,0.16428571428571428,0.1495124593716143
+food_kitchen,56,0.8571428571428571,0.0,0.0,0.8214285714285714,0.0,0.7678571428571429,0.22277227722772278,0.2,0.2107728337236534
+cleaning,8,0.875,0.0,0.0,0.875,0.0,0.625,0.04,0.047619047619047616,0.043478260869565216
+phone_device,162,0.9074074074074074,0.024691358024691357,0.006172839506172839,0.8703703703703703,0.024691358024691357,0.5864197530864198,0.32521739130434785,0.3132328308207705,0.31911262798634815
+paper_cardboard,261,0.9080459770114943,0.034482758620689655,0.011494252873563218,0.8888888888888888,0.034482758620689655,0.7203065134099617,0.22274881516587677,0.32339449541284404,0.2637979420018709
+craft_small_object,106,0.9339622641509434,0.02830188679245283,0.009433962264150943,0.9150943396226415,0.02830188679245283,0.5,0.22662889518413598,0.25806451612903225,0.24132730015082954
+retail_container,101,0.9405940594059405,0.0297029702970297,0.0,0.9108910891089109,0.0297029702970297,0.7722772277227723,0.20279720279720279,0.17522658610271905,0.18800648298217182
+tool_stationery,138,0.9565217391304348,0.014492753623188406,0.0,0.9347826086956522,0.014492753623188406,0.8043478260869565,0.27906976744186046,0.3894523326572008,0.32514817950889074
+no_object_label,2,1.0,0.0,0.0,1.0,0.0,1.0,0.0,0.0,0.0

results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv

@@ -0,0 +1,3 @@
+group,samples,parsed_prediction_rate,action_exact_rate,subtask_exact_rate,transition_exact_rate,next_action_exact_rate,contact_exact_rate,object_precision,object_recall,object_f1
+unseen_in_train,317,0.8454258675078864,0.015772870662460567,0.006309148264984227,0.8233438485804416,0.015772870662460567,0.6151419558359621,0.15804806991988346,0.23183760683760685,0.18796015591165008
+seen_in_train,131,0.9541984732824428,0.04580152671755725,0.007633587786259542,0.916030534351145,0.04580152671755725,0.7175572519083969,0.3185011709601874,0.31627906976744186,0.3173862310385064

scripts/build_artifact_index.py

@@ -129,6 +129,14 @@ ARTIFACTS = [
         "surface": "repo_hf",
         "shows": "Builds synthetic verified packages for every configured backbone and audits them against the public-safe package contract.",
     },
+    {
+        "id": "qwen3_omni_error_analysis_script",
+        "title": "Qwen3-Omni held-out error-analysis script",
+        "path": "scripts/omni/analyze_qwen3_omni_errors.py",
+        "kind": "scaleup_contract",
+        "surface": "repo_hf",
+        "shows": "Computes public-safe held-out error-analysis tables by episode, action family, train-seen status, required-modality state, and object category.",
+    },
     {
         "id": "additional_development_directions",
         "title": "Additional development directions",
@@ -674,6 +682,22 @@ ARTIFACTS = [
         "surface": "repo_hf",
         "shows": "Documents the public multi-episode access status and 32-episode pilot selection.",
     },
+    {
+        "id": "qwen3_omni_error_analysis_report",
+        "title": "Qwen3-Omni held-out error-analysis report",
+        "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+        "kind": "scaleup_status",
+        "surface": "repo_hf",
+        "shows": "Summarizes validation-aware Qwen3-Omni held-out failures by episode, action family, train-seen status, required-modality state, and object category.",
+    },
+    {
+        "id": "qwen3_omni_error_analysis_json",
+        "title": "Qwen3-Omni held-out error-analysis JSON",
+        "path": "results/omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+        "kind": "scaleup_status",
+        "surface": "repo_hf",
+        "shows": "Machine-readable Qwen3-Omni held-out error analysis with grouped metrics and sanitized failure examples.",
+    },
     {
         "id": "citation",
         "title": "Citation metadata",

scripts/omni/analyze_qwen3_omni_errors.py

@@ -0,0 +1,370 @@
+#!/usr/bin/env python3
+"""Analyze public-safe Qwen3-Omni held-out prediction errors.
+
+The script consumes a verified public package, not raw Xperience-10M data. It
+summarizes where the diagnostic pilot fails by episode, train-seen status,
+coarse action family, object category, parsed prediction state, and
+required-modality state. The outputs are small derived CSV/JSON/Markdown
+artifacts suitable for the public package.
+"""
+
+from __future__ import annotations
+
+import argparse
+import csv
+import json
+from collections import Counter, defaultdict
+from pathlib import Path
+from typing import Any
+
+
+DEFAULT_PACKAGE = (
+    Path(__file__).resolve().parents[2]
+    / "results/omni_finetune/verified_public/"
+    / "xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval"
+)
+
+ACTION_FAMILIES = [
+    ("phone_use", ("phone", "smartphone", "watch", "screen")),
+    ("paper_cardboard_craft", ("paper", "cardboard", "fold", "cut", "draw", "mark", "ruler", "scissors", "lantern", "star")),
+    ("retail_stocking", ("shelf", "product", "can", "canned", "container", "box", "grocery", "stock")),
+    ("small_object_sorting", ("bead", "button", "tile", "mahjong", "puzzle", "piece")),
+    ("cleaning", ("clean", "wipe", "wash", "vacuum", "sweep", "trash")),
+    ("locomotion", ("walk", "approach", "enter", "move through", "arrive", "leave")),
+    ("food_kitchen", ("kettle", "rice", "saucepan", "kitchen", "bottle", "jar", "lid")),
+]
+
+OBJECT_CATEGORIES = [
+    ("phone_device", ("phone", "smartphone", "watch", "charger", "cable", "power bank", "earbud")),
+    ("paper_cardboard", ("paper", "cardboard", "lantern", "origami", "star", "ribbon")),
+    ("tool_stationery", ("scissors", "knife", "ruler", "marker", "pen", "stapler", "glue", "tape")),
+    ("retail_container", ("shelf", "container", "product", "box", "can", "canned", "package", "bag")),
+    ("furniture_room", ("table", "chair", "desk", "counter", "sink", "door", "wall", "floor")),
+    ("food_kitchen", ("kettle", "rice", "saucepan", "jar", "bottle", "food", "kitchen")),
+    ("craft_small_object", ("bead", "button", "tile", "mahjong", "puzzle", "foam", "piece")),
+    ("cleaning", ("vacuum", "broom", "cloth", "towel", "trash")),
+]
+
+REQUIRED_VIDEO_FILES = {
+    "fisheye_cam0.mp4",
+    "fisheye_cam1.mp4",
+    "fisheye_cam2.mp4",
+    "fisheye_cam3.mp4",
+    "stereo_left.mp4",
+    "stereo_right.mp4",
+}
+
+REQUIRED_HDF5_MODALITIES = {
+    "calibration",
+    "slam_pose",
+    "slam_point_cloud",
+    "depth",
+    "depth_confidence",
+    "hand_mocap",
+    "body_mocap",
+    "contacts",
+    "imu",
+    "caption",
+}
+
+
+def parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--package-dir", type=Path, default=DEFAULT_PACKAGE)
+    parser.add_argument("--output-dir", type=Path)
+    parser.add_argument("--max-examples", type=int, default=12)
+    return parser.parse_args()
+
+
+def load_json(path: Path) -> dict[str, Any]:
+    return json.loads(path.read_text(encoding="utf-8"))
+
+
+def load_jsonl(path: Path) -> list[dict[str, Any]]:
+    rows = []
+    with path.open("r", encoding="utf-8") as handle:
+        for line in handle:
+            line = line.strip()
+            if line:
+                rows.append(json.loads(line))
+    return rows
+
+
+def norm(value: Any) -> str:
+    return str(value or "").strip().lower()
+
+
+def family_for(text: str, families: list[tuple[str, tuple[str, ...]]], fallback: str = "other") -> str:
+    low = norm(text)
+    for name, keywords in families:
+        if any(keyword in low for keyword in keywords):
+            return name
+    return fallback
+
+
+def object_categories(objects: list[Any]) -> set[str]:
+    categories: set[str] = set()
+    for obj in objects:
+        categories.add(family_for(str(obj), OBJECT_CATEGORIES, "other_object"))
+    return categories or {"no_object_label"}
+
+
+def f1(precision: float, recall: float) -> float:
+    if precision + recall == 0:
+        return 0.0
+    return 2 * precision * recall / (precision + recall)
+
+
+def bool_metric(row: dict[str, Any], key: str) -> bool:
+    true_json = row.get("true_json") or {}
+    pred_json = row.get("pred_json") or {}
+    return norm(true_json.get(key)) == norm(pred_json.get(key)) and bool(pred_json)
+
+
+def object_overlap(row: dict[str, Any]) -> tuple[int, int, int]:
+    true_objects = {norm(item) for item in (row.get("true_json") or {}).get("objects", []) if norm(item)}
+    pred_objects = {norm(item) for item in (row.get("pred_json") or {}).get("objects", []) if norm(item)}
+    return len(true_objects & pred_objects), len(pred_objects), len(true_objects)
+
+
+def modality_state(episode: dict[str, Any] | None) -> tuple[str, list[str]]:
+    if not episode:
+        return "episode_manifest_missing", ["episode_manifest_missing"]
+    missing: list[str] = []
+    files = {str(item.get("name")): bool(item.get("exists")) for item in episode.get("files", [])}
+    for filename in sorted(REQUIRED_VIDEO_FILES):
+        if not files.get(filename):
+            missing.append(filename)
+    hdf5 = episode.get("hdf5_modalities") or {}
+    for modality in sorted(REQUIRED_HDF5_MODALITIES):
+        if not hdf5.get(modality):
+            missing.append(modality)
+    if missing:
+        return "missing_required_modalities", missing
+    if files.get("visualization.rrd") is False:
+        return "rrd_missing_only_required_modalities_present", ["visualization.rrd"]
+    return "required_modalities_present", []
+
+
+def add_row_stats(bucket: dict[str, Any], row: dict[str, Any]) -> None:
+    bucket["samples"] += 1
+    valid = bool(row.get("pred_json"))
+    bucket["parsed_predictions"] += int(valid)
+    bucket["action_exact"] += int(bool_metric(row, "action"))
+    bucket["subtask_exact"] += int(bool_metric(row, "subtask"))
+    bucket["transition_exact"] += int(bool_metric(row, "transition"))
+    bucket["next_action_exact"] += int(bool_metric(row, "next_action"))
+    bucket["contact_exact"] += int(bool_metric(row, "contact"))
+    matched, pred_count, true_count = object_overlap(row)
+    bucket["object_matched"] += matched
+    bucket["object_predicted"] += pred_count
+    bucket["object_true"] += true_count
+
+
+def empty_bucket() -> dict[str, Any]:
+    return {
+        "samples": 0,
+        "parsed_predictions": 0,
+        "action_exact": 0,
+        "subtask_exact": 0,
+        "transition_exact": 0,
+        "next_action_exact": 0,
+        "contact_exact": 0,
+        "object_matched": 0,
+        "object_predicted": 0,
+        "object_true": 0,
+    }
+
+
+def finalize_bucket(name: str, bucket: dict[str, Any]) -> dict[str, Any]:
+    samples = max(int(bucket["samples"]), 1)
+    precision = bucket["object_matched"] / bucket["object_predicted"] if bucket["object_predicted"] else 0.0
+    recall = bucket["object_matched"] / bucket["object_true"] if bucket["object_true"] else 0.0
+    return {
+        "group": name,
+        "samples": bucket["samples"],
+        "parsed_prediction_rate": bucket["parsed_predictions"] / samples,
+        "action_exact_rate": bucket["action_exact"] / samples,
+        "subtask_exact_rate": bucket["subtask_exact"] / samples,
+        "transition_exact_rate": bucket["transition_exact"] / samples,
+        "next_action_exact_rate": bucket["next_action_exact"] / samples,
+        "contact_exact_rate": bucket["contact_exact"] / samples,
+        "object_precision": precision,
+        "object_recall": recall,
+        "object_f1": f1(precision, recall),
+    }
+
+
+def write_csv(path: Path, rows: list[dict[str, Any]]) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    if not rows:
+        path.write_text("", encoding="utf-8")
+        return
+    with path.open("w", encoding="utf-8", newline="") as handle:
+        writer = csv.DictWriter(handle, fieldnames=list(rows[0].keys()), lineterminator="\n")
+        writer.writeheader()
+        writer.writerows(rows)
+
+
+def top_rows(groups: dict[str, dict[str, Any]], *, min_samples: int = 1, reverse: bool = False) -> list[dict[str, Any]]:
+    rows = [finalize_bucket(name, bucket) for name, bucket in groups.items() if bucket["samples"] >= min_samples]
+    return sorted(rows, key=lambda row: (row["parsed_prediction_rate"], row["action_exact_rate"], row["samples"]), reverse=reverse)
+
+
+def markdown_table(rows: list[dict[str, Any]], columns: list[str], limit: int = 8) -> list[str]:
+    selected = rows[:limit]
+    if not selected:
+        return ["No rows."]
+    lines = ["| " + " | ".join(columns) + " |", "| " + " | ".join("---" for _ in columns) + " |"]
+    for row in selected:
+        values = []
+        for col in columns:
+            value = row.get(col)
+            if isinstance(value, float):
+                values.append(f"{value:.4f}")
+            else:
+                values.append(str(value))
+        lines.append("| " + " | ".join(values) + " |")
+    return lines
+
+
+def main() -> int:
+    args = parse_args()
+    package_dir = args.package_dir.expanduser().resolve()
+    output_dir = args.output_dir or package_dir / "analysis"
+    output_dir = output_dir.expanduser().resolve()
+
+    predictions = load_jsonl(package_dir / "eval" / "predictions.jsonl")
+    metrics = load_json(package_dir / "eval" / "metrics.json")
+    episode_manifest = load_json(package_dir / "dataset" / "episode_manifest.json")
+    episodes = {episode.get("episode_id"): episode for episode in episode_manifest.get("episodes", [])}
+
+    overall = empty_bucket()
+    by_episode: dict[str, dict[str, Any]] = defaultdict(empty_bucket)
+    by_family: dict[str, dict[str, Any]] = defaultdict(empty_bucket)
+    by_seen: dict[str, dict[str, Any]] = defaultdict(empty_bucket)
+    by_modality: dict[str, dict[str, Any]] = defaultdict(empty_bucket)
+    by_object_category: dict[str, dict[str, Any]] = defaultdict(empty_bucket)
+    invalid_examples = []
+    overgenerated_examples = []
+    modality_missing_by_episode: dict[str, list[str]] = {}
+
+    for row in predictions:
+        episode_id = str(row.get("episode_id"))
+        true_json = row.get("true_json") or {}
+        pred_json = row.get("pred_json") or {}
+        add_row_stats(overall, row)
+        add_row_stats(by_episode[episode_id], row)
+        add_row_stats(by_family[family_for(str(true_json.get("action")), ACTION_FAMILIES)], row)
+        add_row_stats(by_seen["seen_in_train" if row.get("true_label_seen_in_train") else "unseen_in_train"], row)
+        state, missing = modality_state(episodes.get(episode_id))
+        modality_missing_by_episode.setdefault(episode_id, missing)
+        add_row_stats(by_modality[state], row)
+        for category in object_categories(true_json.get("objects", [])):
+            add_row_stats(by_object_category[category], row)
+        if not pred_json and len(invalid_examples) < args.max_examples:
+            invalid_examples.append({
+                "id": row.get("id"),
+                "episode_id": episode_id,
+                "true_action": true_json.get("action"),
+                "raw_prediction_prefix": str(row.get("raw_prediction", ""))[:240],
+            })
+        pred_objects = pred_json.get("objects", []) if isinstance(pred_json, dict) else []
+        if len(pred_objects) > 20 and len(overgenerated_examples) < args.max_examples:
+            overgenerated_examples.append({
+                "id": row.get("id"),
+                "episode_id": episode_id,
+                "true_action": true_json.get("action"),
+                "predicted_object_count": len(pred_objects),
+                "first_predicted_objects": pred_objects[:20],
+            })
+
+    episode_rows = top_rows(by_episode)
+    family_rows = top_rows(by_family)
+    seen_rows = top_rows(by_seen)
+    modality_rows = top_rows(by_modality)
+    object_rows = top_rows(by_object_category)
+
+    write_csv(output_dir / "episode_error_analysis.csv", episode_rows)
+    write_csv(output_dir / "action_family_error_analysis.csv", family_rows)
+    write_csv(output_dir / "train_seen_error_analysis.csv", seen_rows)
+    write_csv(output_dir / "missing_modality_error_analysis.csv", modality_rows)
+    write_csv(output_dir / "object_category_error_analysis.csv", object_rows)
+
+    summary = {
+        "status": "pass",
+        "source_package": package_dir.name,
+        "source_prediction_rows": len(predictions),
+        "metrics_json_validity_rate": metrics.get("json_validity_rate"),
+        "computed": finalize_bucket("overall", overall),
+        "worst_episode_groups": episode_rows[:8],
+        "action_family_groups": family_rows,
+        "train_seen_groups": seen_rows,
+        "missing_modality_groups": modality_rows,
+        "object_category_groups": object_rows,
+        "invalid_json_examples": invalid_examples,
+        "object_overgeneration_examples": overgenerated_examples,
+        "modality_missing_by_episode": modality_missing_by_episode,
+        "interpretation": (
+            "The diagnostic pilot is dominated by invalid or weak structured outputs and exact-label failures. "
+            "These tables identify where to tighten JSON constraints, action/subtask target formatting, object vocabularies, "
+            "and missing-modality robustness before claiming stronger model quality."
+        ),
+    }
+    (output_dir / "error_analysis_summary.json").write_text(json.dumps(summary, indent=2) + "\n", encoding="utf-8")
+
+    report = [
+        "# Qwen3-Omni Held-Out Error Analysis",
+        "",
+        "This report is computed from the verified public package predictions. It contains only derived metrics and sanitized examples.",
+        "",
+        "## Overall",
+        "",
+        f"- Prediction rows: `{len(predictions)}`",
+        f"- JSON validity from `metrics.json`: `{summary['metrics_json_validity_rate']:.4f}`",
+        f"- Parsed prediction rate from public rows: `{summary['computed']['parsed_prediction_rate']:.4f}`",
+        f"- Action exact rate: `{summary['computed']['action_exact_rate']:.4f}`",
+        f"- Subtask exact rate: `{summary['computed']['subtask_exact_rate']:.4f}`",
+        f"- Contact exact rate: `{summary['computed']['contact_exact_rate']:.4f}`",
+        f"- Object F1: `{summary['computed']['object_f1']:.4f}`",
+        "",
+        "## Weakest Episode Groups",
+        "",
+        *markdown_table(episode_rows, ["group", "samples", "parsed_prediction_rate", "action_exact_rate", "object_f1"]),
+        "",
+        "## Action Families",
+        "",
+        *markdown_table(family_rows, ["group", "samples", "parsed_prediction_rate", "action_exact_rate", "subtask_exact_rate", "object_f1"]),
+        "",
+        "## Train-Seen Split",
+        "",
+        *markdown_table(seen_rows, ["group", "samples", "parsed_prediction_rate", "action_exact_rate", "next_action_exact_rate"]),
+        "",
+        "## Required-Modality State",
+        "",
+        *markdown_table(modality_rows, ["group", "samples", "parsed_prediction_rate", "action_exact_rate", "object_f1"]),
+        "",
+        "## Object Categories",
+        "",
+        *markdown_table(object_rows, ["group", "samples", "object_precision", "object_recall", "object_f1"]),
+        "",
+        "## Interpretation",
+        "",
+        summary["interpretation"],
+        "",
+        "Generated files:",
+        "",
+        "- `error_analysis_summary.json`",
+        "- `episode_error_analysis.csv`",
+        "- `action_family_error_analysis.csv`",
+        "- `train_seen_error_analysis.csv`",
+        "- `missing_modality_error_analysis.csv`",
+        "- `object_category_error_analysis.csv`",
+    ]
+    (output_dir / "ERROR_ANALYSIS.md").write_text("\n".join(report) + "\n", encoding="utf-8")
+    print(json.dumps({"status": "pass", "output_dir": str(output_dir), "prediction_rows": len(predictions)}, indent=2))
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

scripts/validate_mirror_parity.py

@@ -30,6 +30,7 @@ DATA_FILES = [
     "foundation_model_plan.json",
     "live_publication_status.json",
     "modality_atlas.json",
+    "omni_finetune_verified_result.json",
     "project_brief.json",
     "project_manifest.json",
     "project_packet.json",
@@ -76,6 +77,7 @@ ASSET_FILES = [
 ]
 
 SCRIPT_FILES = [
+    "omni/analyze_qwen3_omni_errors.py",
     "audio_ablation_and_raw_upgrade.py",
     "build_artifact_index.py",
     "build_brand_assets.py",
@@ -122,9 +124,18 @@ RESULT_FILES = [
     "single_episode_diagnostics/timeline_overlay/timeline_overlay.csv",
     "single_episode_diagnostics/alignment_stress/alignment_shift_metrics.csv",
     "single_episode_diagnostics/alignment_stress/alignment_stress_summary.json",
+    "omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/ERROR_ANALYSIS.md",
+    "omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/error_analysis_summary.json",
+    "omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/episode_error_analysis.csv",
+    "omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/action_family_error_analysis.csv",
+    "omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/train_seen_error_analysis.csv",
+    "omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/missing_modality_error_analysis.csv",
+    "omni_finetune/verified_public/xperience10m_qwen3_omni_128ep_96train_16val_16test_valmon_20260605_eval/analysis/object_category_error_analysis.csv",
 ]
 
 DOC_FILES = [
+    "ARTIFACT_GUIDE.md",
+    "OMNI_MODEL_EXTENSION_CONTRACT.md",
     "QUALITY_GATES.md",
     "EVALUATION_PROTOCOL.md",
     "FIGURE_INDEX.md",